24 November 2015

🪙Business Intelligence: Discovery (Just the Quotes)

"Put simply, statistics is a range of procedures for gathering, organizing, analyzing and presenting quantitative data. […] Essentially […], statistics is a scientific approach to analyzing numerical data in order to enable us to maximize our interpretation, understanding and use. This means that statistics helps us turn data into information; that is, data that have been interpreted, understood and are useful to the recipient. Put formally, for your project, statistics is the systematic collection and analysis of numerical data, in order to investigate or discover relationships among phenomena so as to explain, predict and control their occurrence." (Reva B Brown & Mark Saunders, "Dealing with Statistics: What You Need to Know", 2008)

"Discovery is the key to building great data products, as opposed to products that are merely good." (Mike Loukides, "The Evolution of Data Products", 2011)

"These practices - selective reporting and data pillaging - are known as data grubbing. The discovery of statistical significance by data grubbing shows little other than the researcher’s endurance. We cannot tell whether a data grubbing marathon demonstrates the validity of a useful theory or the perseverance of a determined researcher until independent tests confirm or refute the finding. But more often than not, the tests stop there. After all, you won’t become a star by confirming other people’s research, so why not spend your time discovering new theories? The data-grubbed theory consequently sits out there, untested and unchallenged." (Gary Smith, "Standard Deviations", 2014)

"Data Lake induces accessibility and catalyzes availability. It warrants data discovery platforms to soak the data trends at a horizontal scale and produce visual insights. It largely cuts down the time that goes into data preparation and exhaustive data analysis." (Saurabh Gupta et al, "Practical Enterprise Data Lake Insights", 2018)

"Analysis is a two-step process that has an exploratory and an explanatory phase. In order to create a powerful data story, you must effectively transition from data discovery (when you’re finding insights) to data communication (when you’re explaining them to an audience). If you don’t properly traverse these two phases, you may end up with something that resembles a data story but doesn’t have the same effect. Yes, it may have numbers, charts, and annotations, but because it’s poorly formed, it won’t achieve the same results." (Brent Dykes, "Effective Data Storytelling: How to Drive Change with Data, Narrative and Visuals", 2019)

"A data product encapsulates more than just the data. It needs to contain all the structural components needed to manifest its baseline usability characteristics - discoverable, understandable, addressable, etc. - in an autonomous fashion, while continuing to share data in a compliant and secure manner." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Overall [...] everyone also has a need to analyze data. The ability to analyze data is vital in its understanding of product launch success. Everyone needs the ability to find trends and patterns in the data and information. Everyone has a need to ‘discover or reveal (something) through detailed examination’, as our definition says. Not everyone needs to be a data scientist, but everyone needs to drive questions and analysis. Everyone needs to dig into the information to be successful with diagnostic analytics. This is one of the biggest keys of data literacy: analyzing data." (Jordan Morrow, "Be Data Literate: The data literacy skills everyone needs to succeed", 2021)

"A data product is based on semantically related raw data that is transformed into a meaningful business context and easily discoverable and consumable by business users." (Eberhard Hechler et al, "Data Fabric and Data Mesh Approaches with AI", 2023)

"The terms Data Fabric and Data Mesh are often viewed as different, conflicting, or at the best overlapping data architectures or frameworks, data management concepts, or approaches to discover, explore, govern, and consume data. However, these concepts are related to each other, where each concept emphasizes specific imperatives or objectives."(Eberhard Hechler et al, "Data Fabric and Data Mesh Approaches with AI", 2023)

"A data mesh is a decentralized data architecture with four specific characteristics. First, it requires independent teams within designated domains to own their analytical data. Second, in a data mesh, data is treated and served as a product to help the data consumer to discover, trust, and utilize it for whatever purpose they like. Third, it relies on automated infrastructure provisioning. And fourth, it uses governance to ensure that all the independent data products are secure and follow global rules." (James Serra, "Deciphering Data Architectures", 2024)

"A data lake is a distributed repository of raw and unprocessed data stored in its original format, without a predefined schema or structure. A data lake is designed to support a wide range of data types, sources, and use cases, such as exploration, discovery, and data experimentation. A data lake follows a 'schema on read' approach. Data is structured and processed only when it is accessed or consumed by a user or application (Extract, Load, Transform (ELT)). A data lake also enables data democratization, meaning data is accessible and available to anyone who needs it, without barriers or restrictions." (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

🪙Business Intelligence: Mistakes (Just the Quotes)

"Using data quality rules brings comprehensive data quality assessment from fantasy world to reality. However, it is by no means simple, and it takes a skillful skipper to navigate through the powerful currents and maelstroms along the way. Considering the volume and structural complexity of a typical database, designing a comprehensive set of data quality rules is a daunting task. The number of rules will often reach hundreds or even thousands. When some rules are missing, the results of the data quality assessment can be completely jeopardized, Thus the first challenge is to design all rules and make sure that they indeed identify all or most errors." (Arkady Maydanchik, "Data Quality Assessment", 2007)

"A common misconception about BI standardization is the assumption that all users must use the same tool. It would be a mistake to pursue this strategy. Instead, successful BI companies use the right tool for the right user. For a senior executive, the right tool might be a dashboard. For a power user, it might be a business query tool. For a call center agent, it might be a custom application or a BI gadget embedded in an operational application."(Cindi Howson, "Successful Business Intelligence: Secrets to making BI a killer App", 2008)

"Business intelligence tools can only present the facts. Removing biases and other errors in decision making are dynamics of company culture that affect how well business intelligence is used." (Cindi Howson, "Successful Business Intelligence: Secrets to making BI a killer App", 2008)

"The problem is when biases and inaccurate data also get filtered into the gut. In this case, the gut-feel decision making should be supported with objective data, or errors in decision making may occur." (Cindi Howson, "Successful Business Intelligence: Secrets to making BI a killer App", 2008

"Tests are sometimes mistaken with quality assurance. These two notions are not identical: 1) quality assurance ensures that the organization's processes are implemented and applied correctly; 2) testing identifies defects and failures, and provides information on the software and the risks associated with their release to the market." (Bernard Homes, "Fundamentals of Software Testing", 2012)

"[…] humans make mistakes when they try to count large numbers in complicated systems. They make even greater errors when they attempt - as they always do - to reduce complicated systems to simple numbers." (Zachary Karabell, "The Leading Indicators: A short history of the numbers that rule our world", 2014)

"Most discussions of decision making assume that only senior executives make decisions or that only senior executives' decisions matter. This is a dangerous mistake. Decisions are made at every level of the organization, beginning with individual professional contributors and frontline supervisors. These apparently low-level decisions are extremely important in a knowledge-based organization." (Zach Gemignani et al, "Data Fluency", 2014)

"Exploratory data analysis is the search for patterns and trends in a given data set. Visualization techniques play an important part in this quest. Looking carefully at your data is important for several reasons, including identifying mistakes in collection/processing, finding violations of statistical assumptions, and suggesting interesting hypotheses." (Steven S Skiena, "The Data Science Design Manual", 2017)

"A common mistake when implementing a data catalog is to focus only on technical metadata. This limits its use and the potential value. It also excludes business users who have valuable related input or need to use the catalog. A catalog should in fact function as a two-way translation layer between technical and business users." (Fadi Maali & Jason Lim, "Implementing a Modern Data Catalog to Power Data Intelligence: Make Trustworthy Data Central to Your Organization", 2022)

"Since data engineering is such a crucial field, you may be wondering who the main players are and what skill sets they possess. Building a data product involves several folks, all of whom need to come together with seamless handoffs to ensure a successful end product or service is created. It would be a mistake to create silos and increase both the number and complexity of integration points as each additional integration is a potential failure point." (Anindita Mahapatra, "Simplifying Data Engineering and Analytics with Delta", 2022)

"The problem of bad data has existed for a very long time. Data copies diverge as their original source changes. Copies get stale. Errors detected in one data set are not fixed in duplicate ones. Domain knowledge related to interpreting and understanding data remains incomplete, as does support from the owners of the original data." (Adam Bellemare, "Building an Event-Driven Data Mesh: Patterns for Designing and Building Event-Driven Architectures", 2023)

"Domain autonomy should not be mistaken for a lack of governance or accountability. Autonomy, in this context, implies a higher level of responsibility. Domains are free to act and accountable for their actions, especially regarding how well their data strategies align with domain-specific and broader organizational objectives." (Pradeep Menon, "Data Mesh Principles, patterns, architecture, and strategies for data-driven decision making", 2024)

23 November 2015

♟️Strategic Management: Methods (Just the Quotes)

"The writer has found, in analyzing and diagnosing organization and accounting work, that charts can express more on one page than is sometimes expressed in several chapters of writing, and has been the author and originator of many methods of charting industrial expressions. It is necessary, as a first step, for analytical and other purposes, to make a chart expressing all of the relations governing the organization of a business so as to show the very foundation upon which all authorities, accounting, and business transactions are based and conducted. There have been more failures scored both personally and financially for lack of these very elements in a business than by reason of any other one thing. As well try to build a house without a foundation as to try to conduct a business, especially a manufacturing business, without proper organization." (Clinton E. Woods, "Organizing a factory", 1905)

"It is only through enforced standardization of methods, enforced adoption of the best implements and working conditions, and enforced cooperation that this faster work can be assured. And the duty of enforcing the adoption of standards and enforcing this cooperation rests with management alone." (Frederick W Taylor, "Principles of Scientific Management", 1911)

"Motion study is the science of eliminating wastefulness resulting from using unnecessary, ill-directed, and inefficient motions. The aim of motion study is to find and perpetuate the scheme of least waste methods of labor." (Frank B Gilbreth, "Primer of scientific management", 1912) 

"For any manager to utilize graphic methods for visualizing the vital facts of his business, in the first place it must be impressed upon his that the method will produce the results for him and then he must know how to get up a chart correctly, and last, but far from least, he must know what the essential facts of his business are. Charts, in themselves, mean little and like many another force for the accomplishment of good, if misdirected, may result unprofitably." (Allan C Haskell, "How to Make and Use Graphic Charts", 1919)

"Business executives cannot afford to ignore the merits of graphical representation which have for so long been accepted by the engineer and man of science. They must look behind the graphical method and study the conditions leading to the picture along with the picture itself. No business is too small to profit by an examination which shall analyze and scrutinize nor too large to ignore its possibilities. Each business must adjust the graphical methods to its own peculiarities and each diagram must be adjusted to the individual for whom it is prepared or the individual must be educated up to the use and importance of these methods of analysis." (William C Marshall, "Graphical methods for schools, colleges, statisticians, engineers and executives", 1921)

"Do not confuse objectives with methods. When the nation becomes substantially united in favor of planning the broad objectives of civilization, then true leadership must unite thought behind definite methods." (Franklin D Roosevelt, 1937)

"When an active individual of sound common sense perceives the sordid state of the world, desire to change it becomes the guiding principle by which he organizes given facts and shapes them into a theory. The methods and categories as well as the transformation of the theory can be understood only in connection with his taking of sides. This, in turn, discloses both his sound common sense and the character of the world. Right thinking depends as much on right willing as right willing on right thinking." (Max Horkheimer, "The Latest Attack on Metaphysics", 1937)

"The concern of OR with finding an optimum decision, policy, or design is one of its essential characteristics. It does not seek merely to define a better solution to a problem than the one in use; it seeks the best solution... [It] can be characterized as the application of scientific methods, techniques, and tools to problems involving the operations of systems so as to provide those in control of the operations with optimum solutions to the problems." (C West Churchman et al, "Introduction to Operations Research", 1957)

"Linking the basic parts are communication, balance or system parts maintained in harmonious relationship with each other and decision making. The system theory include both man-machine and interpersonal relationships. Goals, man, machine, method, and process are woven together into a dynamic unity which reacts." (George R Terry, "Principles of Management", 1960)

"The essential task of management is to arrange organizational conditions and methods of operations so that people can achieve their own goals best by directing their own efforts toward organizational objectives." (Douglas McGregor, "The Human Side of Enterprise", 1960)

"The unique feature of the decision tree is that it allows management to combine analytical techniques such as discounted cash flow and present value methods with a clear portrayal of the impact of future decision alternatives and events. Using the decision tree, management can consider various courses of action with greater ease and clarity. The interactions between present decision alternatives, uncertain events, and future choices and their results become more visible." (John F Magee, "Decision Trees for Decision Making", Harvard Business Review, 1964)

"The concept of leadership has an ambiguous status in organizational practice, as it does in organizational theory. In practice, management appears to be of two minds about the exercise of leadership. Many jobs are so specified in content and method that within very broad limits differences among individuals become irrelevant, and acts of leadership are regarded as gratuitous at best, and at worst insubordinate." (Daniel Katz & Robert L Kahn, "The Social Psychology of Organizations", 1966)

"For the scientist a model is also a way in which the human though processes can be amplified. This method often takes the form of models that can be programmed into computers. At no point, however, the scientist intend to loose control of the situation because off the computer does some of his thinking for him. The scientist controls the basic assumptions and the computer only derives some of the more complicated implications." (C West Churchman, "The Systems Approach", 1968)

"It is an axiom of program budgeting that the budget should facilitate the process of alternative methods of obtaining objectives." (Chester Wright, "Program Budgeting and Cost Benefit Analysis", 1969)

"Statistics is a body of methods and theory applied to numerical evidence in making decisions in the face of uncertainty." (Lawrence Lapin, "Statistics for Modern Business Decisions", 1973)

"Strategic planning is not the 'application of scientific methods to business decision' […] . It is the application of thought, analysis, imagination, and judgment. It is responsibility, rather than technique. […] Strategy planning is not forecasting. […] Strategic planning is necessary precisely because we cannot forecast. […] Strategic planning does nor deal with future decisions. It deals with the futurity of present decisions. […] Strategic planning is not an attempt to eliminate risk. It is not even an attempt to minimize risk." (Peter F Drucker, "Management: Tasks, Responsibilities, Practices", 1973)

"Perhaps the fault [for the poor implementation record for models] lies in the origins of managerial model-making - the translation of methods and principles of the physical sciences into wartime operations research. [...] If hypothesis, data, and analysis lead to proof and new knowledge in science, shouldn’t similar processes lead to change in organizations? The answer is obvious-NO! Organizational changes" (or decisions or policies) do not instantly pow from evidence, deductive logic, and mathematical optimization." (Edward B Roberts, "Interface", 1977)

"Someone adhering to the values of a corporate culture - an intelligent corporate citizen - will behave in consistent fashion under similar conditions, which means that managers don’t have to suffer the inefficiencies engendered by formal rules, procedures, and regulations. […] management has to develop and nurture the common set of values, objectives, and methods essential to the existence of trust. How do we do that? One way is by articulation, by spelling [them] out. […] The other even more important way is by example." (Andrew S Grove, "High Output Management", 1983)

"The formal structure of a decision problem in any area can be put into four parts:" (1) the choice of an objective function denning the relative desirability of different outcomes;" (2) specification of the policy alternatives which are available to the agent, or decisionmaker," (3) specification of the model, that is, empirical relations that link the objective function, or the variables that enter into it, with the policy alternatives and possibly other variables; and" (4) computational methods for choosing among the policy alternatives that one which performs best as measured by the objective function." (Kenneth Arrow, "The Economics of Information", 1984)

"A real challenge for some organizations is to build more qualitative information into their formal systems. One method used in some companies is to request a written narrative with each submission of statistics from the field. Another method is to hold periodic, indepth discussions involving several managers from different levels so that each can contribute whatever qualitative data are available to him." (Larry E Greiner et al, "Human Relations", 1986)

"Enterprise Engineering is not a single methodology, but a sophisticated synthesis of the most important and successful of today's change methods. 'Enterprise Engineering' first explains in detail all the critical disciplines (including continuous improvement, radical reinvention of business processes, enterprise redesign, and strategic visioning). It then illustrates how to custom-design the right combination of these change methods for your organization's specific needs." (James Martin, "The Great Transition, 1995)

"Enterprise Engineering is defined as that body of knowledge, principles, and practices having to do with the analysis, design, implementation and operation of an enterprise. In a continually changing and unpredictable competitive environment, the Enterprise Engineer addresses a fundamental question: 'how to design and improve all elements associated with the total enterprise through the use of engineering and analysis methods and tools to more effectively achieve its goals and objectives' [...]" (Donald H Liles, "The Enterprise Engineering Discipline", 1996)

"Change pressures arise from different sectors of a system. At times it is mandated from the top of a hierarchy, other times it forms from participants at a grass-roots level. Some changes are absorbed by the organization without significant impact on, or alterations of, existing methods. In other cases, change takes root. It causes the formation of new methods" (how things are done and what is possible) within the organization." (George Siemens, "Knowing Knowledge", 2006)

"It's not enough to be talented. It's not enough to work hard and to study late into the night. You must also become intimately aware of the methods you use to reach your decisions." (Garry Kasparov, "How Life Imitates Chess", 2007)

"The goal of enterprise architecture is to create a unified IT environment" (standardized hardware and software systems) across the firm or all of the firm's business units, with tight symbiotic links to the business side of the organization" (which typically is 90% of the firm […] at least by way of budget). More specifically, the goals are to promote alignment, standardization, reuse of existing IT assets, and the sharing of common methods for project management and software development across the organization." (Daniel Minoli, "Enterprise architecture A to Z: frameworks, business process modeling", 2008)

"Enterprise architecture [is] a coherent whole of principles, methods, and models that are used in the design and realisation of an enterprise's organisational structure, business processes, information systems, and infrastructure. […] The most important characteristic of an enterprise architecture is that it provides a holistic view of the enterprise. […] To achieve this quality in enterprise architecture, bringing together information from formerly unrelated domains necessitates an approach that is understood by all those involved from those different domains." (Marc Lankhorst, "Enterprise Architecture at Work: Modelling, Communication and Analysis", 2009)


21 November 2015

🪙Business Intelligence: Schemas (Just the Quotes)

"A data warehouse follows a pre-built static structure to model source data. Any changes at the structural and configuration level must go through a stringent business review process and impact analysis. Data lakes are very agile. Consumption or analytical layer can be modified to fit in the model requirements. Consumers of a data lake are not constant; therefore, schema and modeling lies at the liberty of analysts and scientists." (Saurabh Gupta et al, "Practical Enterprise Data Lake Insights", 2018)

"Data management of the future must build in embracing change, by default. Rigid data modeling and querying languages that expect to put the system in a straitjacket of a never-changing schema can only result in a fragile and unusable analytics system. [...] The data management of the future must support managing and accessing data across multiple hosting platforms, by default." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"The main challenges include relentlessly chasing data issues that include schema and quality changes (data drift). Sometimes, fixing these issues can cause outages and delays to existing jobs. This is tied tightly to the underlying infrastructure, process, and technology and can be vulnerable to any changes there. For example, a temporary glitch in the cloud ecosystem will result in a failure of the data pipeline." (Anindita Mahapatra, "Simplifying Data Engineering and Analytics with Delta", 2022)

"Data governance creates access controls between the data product producer and consumer and provides metadata like schema definitions and lineages. In some cases, mastered data along with reference data may be relevant to the implementation. Data governance allows us to create appropriate access controls for these resources as well." (Hubert Dulay & Stephen Mooney, "Streaming Data Mesh", 2023)

"Data products should remain stable and be decoupled from the operational/transactional applications. This requires a mechanism for detecting schema drift, and avoiding disruptive changes. It also requires versioning and, in some cases, independent pipelines to run in parallel, giving your data consumers time to migrate from one version to another." (Piethein Strengholt, "Data Management at Scale: Modern Data Architecture with Data Mesh and Data Fabric" 2nd Ed., 2023)

"Enforcing a schema at read time, instead of at write time, leads to a proliferation of what we call 'bad data'. The lack of write-time checks means that data written into HDFS may not adhere to the schemas that the readers are using in their existing work […]. Some bad data will cause consumers to halt processing, while other bad data may go silently undetected. While both of these are problematic, silent failures can be deadly and difficult to detect." (Adam Bellemare, "Building an Event-Driven Data Mesh: Patterns for Designing and Building Event-Driven Architectures", 2023)

"Modern data platforms leveraging ETL for analytics will always be consumers of data as they ingest data from various data sources. And as organizations continue to collect, process, and analyze data from a growing number of data sources, the ability to swiftly handle schema evolution and data validation is a critical aspect of any data platform. In this chapter you have seen how Delta Lake gives you flexibility to evolve a table’s schema through dynamic and explicit schema updates, while also enforcing schema validation." (Bennie Haelen & Dan Davis, "Delta Lake: Up and Running - Modern Data Lakehouse Architectures with Delta Lake", 2023)

"Schema evolution in Delta Lake refers to the ability to evolve the schema of a Delta table over time, while preserving the existing data in the table. In other words, schema evolution allows us to add, remove, or modify columns in an existing Delta table without losing any data or breaking any downstream jobs that depend on the table. This is important as your data and business needs change over time and you may need to add new columns to your table or modify the existing columns to support new use cases." (Bennie Haelen & Dan Davis, "Delta Lake: Up and Running - Modern Data Lakehouse Architectures with Delta Lake", 2023)

"Traditionally, data lakes have always operated under the principle of schema on read, but have always had challenges enforcing schema on write. This means there is no predefined schema when data is written to storage, and a schema is only adapted when the data is processed. It is imperative for the case of analytics and data platforms that your table formats enforce the schema on write to prevent introducing change-breaking processes, and to maintain proper data quality and integrity." (Bennie Haelen & Dan Davis, "Delta Lake: Up and Running - Modern Data Lakehouse Architectures with Delta Lake", 2023)

"A Data Lakehouse provides a unified platform for various data workloads, such as descriptive, predictive, and prescriptive analytics. It can handle structured and unstructured data and enforce schema at both read and write times, enabling traditional business intelligence tasks and advanced analytics on the same platform." (Pradeep Menon, "Data Mesh Principles, patterns, architecture, and strategies for data-driven decision making", 2024)

"A data lake is a distributed repository of raw and unprocessed data stored in its original format, without a predefined schema or structure. A data lake is designed to support a wide range of data types, sources, and use cases, such as exploration, discovery, and data experimentation. A data lake follows a 'schema on read' approach. Data is structured and processed only when it is accessed or consumed by a user or application (Extract, Load, Transform (ELT)). A data lake also enables data democratization, meaning data is accessible and available to anyone who needs it, without barriers or restrictions." (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

"At its core, observability rests on three fundamental pillars: metrics, logs, and traces. In the context of data systems, these pillars translate into quantitative measurements (such as data volume, processing latency, and schema changes), detailed event records (including data pipeline execution logs and error messages), and lineage traces that map the flow of data through interconnected processes. Together, they enable a granular and multidimensional understanding of data system behavior, facilitating not just detection but also rapid root-cause analysis." (William Smith, "Soda Core for Modern Data Quality and Observability: The Complete Guide for Developers and Engineers", 2025)

"Data Lakes embrace a schema-on-read approach, storing vast volumes of raw or lightly processed data in native formats with minimal upfront constraints. This design significantly enhances ingestion velocity and accommodates diverse, unstructured, or semi-structured datasets. However, enforcing data quality at scale becomes more complex, as traditional static constraints are absent." (William Smith, "Great Expectations for Modern Data Quality: The Complete Guide for Developers and Engineers", 2025) 

"For a metadata-driven approach to work, a metastore is crucial. A metastore is a (centralized) repository that stores metadata such as schema structures, source and target locations, ownership information, and usage details. The metastore also helps enforce data policies and rules, as metadata can be used to define and apply these policies project-wide. Additionally, a metastore facilitates sharing metadata across various tools and systems used in data management and engineering." (Piethein Strengholt, "Building Medallion Architectures: Designing with Delta Lake and Spark", 2025)

"The hub and spoke, or 'star network', is a data architecture model that centralizes data from various sources into a single hub, such as a data warehouse or data lake. The hub serves as the source of truth for data and provides standardized schemas and formats. The spokes are the various applications or services that consume data from the hub for different purposes, such as analytics, reporting, or ma-chine learning. Spokes can also perform transformations or aggregations on data before presenting it to end users. The hub and spoke architecture aims to simplify data integration and management by reducing complexity and redundancy in data pipelines" (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

20 November 2015

🪙Business Intelligence: Medallion Architecture (Just the Quotes)

"When you leverage Delta Lake with Structured Streaming, you get both the transactional guarantees of Delta Lake and the powerful programming model of Apache Spark Structured Streaming. With Delta Lake, you can now use Delta tables as both streaming sources and sinks, enabling a continuous processing model that processes your data through the Raw, Bronze, Silver, and Gold data lake layers in a streaming fashion, eliminating the need for batch jobs, resulting in a simplified solution architecture." (Bennie Haelen & Dan Davis, "Delta Lake: Up and Running - Modern Data Lakehouse Architectures with Delta Lake", 2023)

"A lake based on the medallion architecture combines the best of lakes and data warehouses. By breaking down silos and eliminating data duplication, it becomes a standard for building data platform architecture." (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

"A Medallion architecture is a data design pattern used to logically organize data, most often in a lakehouse, using three layers for the data platform, with the goal of incrementally and progressively improving the structure and quality of data as it flows through each layer of the data architecture (from Bronze ⇒ Silver ⇒ Gold layer)." (Piethein Strengholt, "Building Medallion Architectures: Designing with Delta Lake and Spark", 2025)

"But what makes the Medallion architecture so crucial for your organization’s data strategy? And what compelling conclusions can be drawn from this journey? The answer lies in its flexible, modular approach that allows organizations to tailor their data processes to specific needs. While the concept of three distinct layers offers a structured approach, it’s not a one-size-fits-all solution. The key is understanding the strengths and limitations of each layer, which can be adapted to better align with operational realities and strategic goals." (Piethein Strengholt, "Building Medallion Architectures: Designing with Delta Lake and Spark", 2025)

"Implementing a Medallion architecture isn’t about following a checklist set of instructions. It’s not about taking the exact, fixed process and trying to fit your unique organization around it. Instead, focus on providing your data consumers with context -  organize your data so they understand when it has been cleaned, when it is ready for consumption. Help them find the data to empower their work as fast and easily as possible. Once you have that, you can fit the processes to your design." (Piethein Strengholt, "Building Medallion Architectures: Designing with Delta Lake and Spark", 2025)

"[...] Medallion architectures are a fantastic concept that is widely misunderstood. Too often, they’re treated as a rigid, step-by-step framework, when they are actually a flexible approach to making sense of an evolving landscape. They are an attempt to simplify a decade of organic evolution and technical innovation into concepts that can be presented to nontechnical users. But simplification comes at a cost: it leaves huge gaps for debate, misinterpretation, and frustration." (Piethein Strengholt, "Building Medallion Architectures: Designing with Delta Lake and Spark", 2025)

"Medallion architectures emerged as the bridge to guide organizations through the lakehouse era. Instead of being the domain of niche companies with unique technical challenges, data lakes have become the de facto technology for data platforms; the doors have been thrown wide open - come on in, the water’s lovely." (Piethein Strengholt, "Building Medallion Architectures: Designing with Delta Lake and Spark", 2025)

"Traditionally, data engineers are responsible for the first steps of data transformation, commonly referred to as the transition from the 'bronze' stage to the 'silver' stage. This phase includes the normalization of raw data to clean and organize it into a structured and accessible format. Data Engineers ensure that data is properly ingested, stored, and prepared for subsequent steps. Their work focuses on building robust data pipelines and applying basic transformations that make the data usable. Next, responsibility may be handed over to an analytics engineer, who takes charge of the transition from the 'silver' stage to the 'gold' stage. This step involves more complex transformations aimed at refining, enriching, and modeling the data to meet specific analytical needs. The analytics engineer ensures that the data is ready to be used in reports, dashboards, and advanced analyses. The transition to the 'gold' stage means that the data is fully prepared for analytic use, providing strategic insights from consolidated data sources." (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

🪙Business Intelligence: Knowledge Graphs (Just the Quotes)

"Knowledge graphs are a specific type of graph with an emphasis on contextual understanding. Knowledge graphs are interlinked sets of facts that describe real-world entities, events, or things and their interrelations in a human- and machine-understandable format." (Jesús Barrasa et al, "Knowledge Graphs: Data in Context for Responsive Businesses", 2021)

"[…] knowledge graphs are useful because they provide contextualized understanding of data. They achieve this by adding a layer of metadata that imposes rules for structure and interpretation." (Jesús Barrasa et al, "Knowledge Graphs: Data in Context for Responsive Businesses", 2021)

"Knowledge graphs use an organizing principle so that a user" (or a computer system) can reason about the underlying data. The organizing principle gives us an additional layer of organizing data (metadata) that adds connected context to support reasoning and knowledge discovery. […] Importantly, some processing can be done without knowledge of the domain, just by leveraging the features of the property graph model" (the organizing principle)." (Jesús Barrasa et al,Knowledge Graphs: Data in Context for Responsive Businesses", 2021)

"Data Fabric architecture utilizes active metadata, knowledge graphs, and semantic enrichment, combining intelligent information integration and transformation technologies to intelligently support data consumers, for example, business users."  (Eberhard Hechler et al, "Data Fabric and Data Mesh Approaches with AI", 2023)

"In Exploiting semantic knowledge graphs can support interpretability and explainability of nearly all AI model types (including DL models) by discovering and depicting semantic and non-obvious relationships or depicting an ML model in a simplified and more readable, explainable way., a Data Mesh solution organizes data around business domain owners and transforms relevant data assets (data sources) to data products that can be consumed by distributed business users from various business domains or functions. These data products are created, governed, and used in an autonomous, decentralized, and self-service manner. Self-service capabilities, which we have already referenced as a Data Fabric capability, enable business organizations to entertain a data marketplace with shopping-for-data characteristics." (Eberhard Hechler et al, "Data Fabric and Data Mesh Approaches with AI", 2023)

"It is essential to realize that the Data Fabric architecture enables the Data Mesh solution via its rich knowledge catalog, semantic search and discovery, smart integration capabilities, and semantic knowledge graphs. Trustworthy AI, for instance, is enabled via the Data Fabric as well." (Eberhard Hechler et al, "Data Fabric and Data Mesh Approaches with AI", 2023)

"As with many other deep learning-based approaches, another major challenge is in interpretability. While knowledge graphs provide a structured and transparent way to store relationships, LLMs operate as a black box, making it difficult to understand how specific outputs are generated. [...] Data alignment is also a key issue, as structured knowledge graphs and unstructured text data must be carefully preprocessed to ensure consistency.  Differences in data formats, ontology mismatches, and information redundancy can create inefficiencies when integrating these two paradigms. Developing robust pipelines that seamlessly connect graph-based insights with LLM-generated text remains an open challenge." (Aldo Marzullo et al, "Graph Machine Learning" 2nd Ed., 2025)

"Despite their impressive capabilities, LLMs are not without limitations. One of the most significant challenges is the problem of hallucination, where an LLM generates factually incorrect or misleading information that appears plausible. This is particularly problematic in domains requiring high factual accuracy, such as healthcare, finance, and legal applications. To mitigate hallucinations and enhance the reliability of LLM outputs,  Retrieval-Augmented Generation (RAG) has emerged as a powerful technique. RAG works by dynamically retrieving relevant information from an external knowledge source (such as a knowledge graph) at inference time, rather than just relying on pre-trained knowledge. This approach ensures that the model has access to up-to-date and accurate data, grounding answers in verified information rather than generating content purely from its internal representations." (Aldo Marzullo et al, "Graph Machine Learning" 2nd Ed., 2025)

"Despite their effectiveness and advantages in supporting the development of intelligent systems, KGs haven’t been widely adopted for several reasons, including the following: (•) They are expensive to build and maintain in terms of time, effort, and money. (•) Intricate access patterns are required to navigate multiple hops. (•) Their results scatter information across multiple nodes and relationships." (Alessandro Negro, "Graph-Powered Machine Learning", 2021)

"Generative artificial intelligence (GenAI), powered by large language models (LLMs) like Google’s Gemini and OpenAI’s GPT, has transformed how we work and live, revolutionizing business after business. Despite this success, generative AI falls short in domains where specific domain knowledge, high accuracy, and explainability are essential. And it has other significant limitations, including hallucinations and a lack of context and relations. This is where knowledge graphs (KGs) come in, provid-ing contextual information - such as experiences, environmental characteristics, cultural aspects, and social normsneeded to build the 'third wave of AI' for mission-critical applications." (Alessandro Negro et al, "Knowledge Graphs and LLMs in Action", 2026)

"KGs are sophisticated graph structures that represent real-world entities (people, places, diseases, proteins), define meaningful connections between them, and provide context. KGs provide structured, explainable knowledge representation but are challenging to build and query; LLMs offer natural language processing capabilities but suffer from hallucinations, stale information, and a lack of domain-specific grounding. Together, they are a 'killer combination': LLMs can extract entities and relationships from unstructured text to build KGs more efficiently, providing more autonomous and powerful graph querying and analysis. Meanwhile, KGs provide reliable, up-to-date domain knowledge to ground LLM responses and prevent hallucinations." (Alessandro Negro et al, "Knowledge Graphs and LLMs in Action", 2026)

"The KG serves as the central reference for all structured and unstructured data related to a domain. Because a KG represents information by focusing on the meaning of data, users can overcome challenges related to data types, formats, and provenance, connecting information from multiple data sources. [...] A KG represents the core information and big picture of a domain. Humans can analyze, visualize, and query graph data to extract insights. Inference rules and machine learning algorithms are performed on top of the KG to infer new information not explicitly encoded within the KG. Analysts can use methods such as centrality and connectivity analysis to identify influential nodes, network analysis to detect the shortest path between nodes, and community analysis to recognize groups of similar nodes." (Alessandro Negro et al, "Knowledge Graphs and LLMs in Action", 2026) 

"Traditional paradigms build systems for specific purposes with structured, homogeneous databases. This approach works for tailored needs but is impractical for complex domains that need to adapt to user characteristics and integrate heterogeneous data. KGs capture connections, enabling relationship discovery through graph pattern matching and traversal. Both the Resource Description Framework (RDF) and Labeled Property Graphs (LPGs) provide machine-readable formats that humans can interpret. KGs emphasize rich, meaningful data representations usable by both humans and machines, enabling a paradigm shift where intelligent behavior is encoded in a unique source of truth." (Alessandro Negro et al, "Knowledge Graphs and LLMs in Action", 2026)

19 November 2015

🪙Business Intelligence: Data Sharing (Just the Quotes)

"Having multiple data lakes replicates the same problems that were created with multiple data warehouses - disparate data siloes and data fiefdoms that don't facilitate sharing of the corporate data assets across the organization. Organizations need to have a single data lake from which they can source the data for their BI/data warehousing and analytic needs. The data lake may never become the 'single version of the truth' for the organization, but then again, neither will the data warehouse. Instead, the data lake becomes the 'single or central repository for all the organization's data' from which all the organization's reporting and analytic needs are sourced." (Billl Schmarzo, "Driving Business Strategies with Data Science: Big Data MBA" 1st Ed., 2015)

"Data governance policies must not enforce constraints on data - Data governance intends to control the level of democracy within the data lake. Its sole purpose of existence is to maintain the quality level through audits, compliance, and timely checks. Data flow, either by its size or quality, must not be constrained through governance norms. [...] Effective data governance elevates confidence in data lake quality and stability, which is a critical factor to data lake success story. Data compliance, data sharing, risk and privacy evaluation, access management, and data security are all factors that impact regulation." (Saurabh Gupta et al, "Practical Enterprise Data Lake Insights", 2018)

"A data product encapsulates more than just the data. It needs to contain all the structural components needed to manifest its baseline usability characteristics - discoverable, understandable, addressable, etc. - in an autonomous fashion, while continuing to share data in a compliant and secure manner." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Data Mesh is a sociotechnical approach to share, access and manage analytical data in complex and large-scale environments - within or across organizations." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"In the case of data mesh, a data product is an architectural quantum. It is the smallest unit of architecture that can be independently deployed and managed. It has high functional cohesion, i.e., performing a specific analytical transformation and securely sharing the result as domain-oriented analytical data. It has all the structural components that it requires to do its function: the transformation code, the data, the metadata, the policies that govern the data, and its dependencies to infrastructure." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"A data sharing data service shares data, in any format and any size, from multiple sources within an organization or other organizations. This type of service provides the required control to share data and allows data-sharing policies to be created. It also enables data sharing in a structured manner and offers complete visibility into how the data is shared and how it is used. A data-sharing system uses APIs for data sharing." (Pradeep Menon, "Data Lakehouse in Action", 2022)

"An API is an interface that allows applications to interact with an external service using a simple set of commands. Data can also be served as part of API interaction. As the data is exposed to multiple external services, API-based methods can scale to share data securely with external services. Data through an API is served in JSON format, therefore the technology used to serve the data using APIs should be able to support JSON formats. For example, a NoSQL database can store such data." (Pradeep Menon, "Data Lakehouse in Action", 2022)

"Each domain data lakehouse may opt to have its data catalog. However, the critical component in this architecture is the data mesh catalog. The data mesh catalog is the master catalog used to discover the data elements available in different nodes. Each domain-oriented node will donate its metadata to the data mesh catalog. This donation of metadata determines the effectiveness of the data mesh architecture. Once the metadata is contributed, other nodes can browse through the data mesh catalog. They can select the data of interest and mutually share data between the nodes through a governed data sharing process. The critical point to note here is that, unlike the hub-spoke architecture, the data mesh architecture enables data sharing between the 'spoke nodes'. There is no hub node in a data mesh architecture." (Pradeep Menon, "Data Lakehouse in Action", 2022)

"A data strategy must align with the business goals and overall framework of how data will be used and managed within an organization. It needs to include standards for how data will be discovered, integrated, accessed, shared, and protected. It needs to address how data will meet regulatory compliance policies, Master Data Management, and data democratization. There needs to be an assurance that both data and metadata have a quality control framework in place to achieve data trust. A data strategy needs to have a clear path on how an organization will accomplish data monetization." (Sonia Mezzetta, "Principles of Data Fabric: Become a data-driven organization by implementing Data Fabric solutions efficiently", 2023)

"Data lineage is the path the data took from its source origin, the stops it made along the way, and its destination. This includes information on all the systems it passed through, how it was cleansed, what it was enriched with, and how it was secured. Capturing all that metadata is difficult because many of those systems and applications don’t share information. It’s up to you to assemble the data’s path by pulling metadata from all those systems/applications and assembling them in hopes that you find the path your data took from its current location (destination) to its source system. Lineage is probably the hardest piece of metadata to acquire for either streaming or batching data pipelines." (Hubert Dulay & Stephen Mooney, "Streaming Data Mesh", 2023)

"Manage data as a strategic asset that evolves into a data product. The premise here is to stop managing data as a byproduct and create an ecosystem that manages data as a valuable strategic asset that can evolve into a data product. Data producers are accountable for managing the life cycle of data from creation to end of life and ensuring it creates business value along the way for data consumers. This requires data that is governed, trusted, protected, secure, and easily accessible. Move data from technical data assets to Data Products by operationalizing data for high scale sharing." (Sonia Mezzetta, "Principles of Data Fabric: Become a data-driven organization by implementing Data Fabric solutions efficiently", 2023)

"Since domains are used to create data products, and sharing data products across many domains ultimately builds a mesh of data, we need to ensure that the data being served follows some guidelines. Data governance involves creating and adhering to a set of global rules, standards, and policies applied to all data products and their interfaces to ensure a collaborative and interoperable data mesh community. These guidelines must be agreed upon among the participating data mesh domains." (Hubert Dulay & Stephen Mooney, "Streaming Data Mesh", 2023)

"Federation is about providing autonomy to each data product owner to make their own decisions about the storage, computing, and sharing of data. However, this autonomy cannot come at a risk to the security and compliance standards of the company." (Aniruddha Deswandikar, "Engineering Data Mesh in Azure Cloud", 2024)

"When data is considered a product, it creates opportunities for collaboration across different domains. This collaboration involves working with other teams to create, share, and use data products that span multiple areas of expertise, interest, or value. Data Mesh promotes cross-domain collaboration by focusing on the consumers rather than the producers. Data products are made available through standardized interfaces and protocols that support various modes of consumption and are governed by domain experts who understand the context and nuances of their data." (Pradeep Menon, "Data Mesh Principles, patterns, architecture, and strategies for data-driven decision making", 2024)

"A lakehouse is a data storage space that hosts and manages all types of data in one place (structured, semi-struc-tured, and unstructured), allowing different tools to normalize and examine this data according to organizational requirements and/or individual choices. A lakehouse thus combines the best aspects of a data lake and a data warehouse by eliminating data duplication and friction related to ingestion, transformation, and sharing of data within the organization, all in the open format, Delta Lake." (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

"Fabric integrates the various technologies needed for an end-to-end data project (namely, ingestion, preparation, storage, processing, enrichment, analysis, visualization, and data sharing) within a single platform accessible as Software as a Service (SaaS), meaning via a simple connection on a web browser. This reduces complexity, costs, and delays related to using multiple tools and technologies, and eliminates all the operational maintenance of infrastructure serving data analytics needs." (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

"This transition to OneDrive highlights the importance of governance adapted to new methods of collaborative work and data sharing. The idea of OneLake is, therefore, based on this same concept: rather than subscribing to a data lake technology that must be maintained, why not simply subscribe to a storage service that offers a layer of abstraction over the complexities of these data storage infrastructures? As a result, the data lake becomes a controlled or governed environment, but still accessible to users who can view it as a simple and intuitive way to securely share data with their colleagues and IT teams."(Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

🪙Business Intelligence: Context (Just the Quotes)

"Data in isolation are meaningless, a collection of numbers. Only in context of a theory do they assume significance […]" (George Greenstein, "Frozen Star", 1983)

"Without meaningful data there can be no meaningful analysis. The interpretation of any data set must be based upon the context of those data." (Donald J Wheeler, "Understanding Variation: The Key to Managing Chaos" 2nd Ed., 2000)

"Statistical analysis of data can only be performed within the context of selected assumptions, models, and/or prior distributions. A statistical analysis is actually the extraction of substantive information from data and assumptions. And herein lies the rub, understood well by Disraeli and others skeptical of our work: For given data, an analysis can usually be selected which will result in 'information' more favorable to the owner of the analysis then is objectively warranted." (Stephen B Vardeman & Max D Morris, "Statistics and Ethics: Some Advice for Young Statisticians", The American Statistician vol 57, 2003)

"Readability in visualization helps people interpret data and make conclusions about what the data has to say. Embed charts in reports or surround them with text, and you can explain results in detail. However, take a visualization out of a report or disconnect it from text that provides context (as is common when people share graphics online), and the data might lose its meaning; or worse, others might misinterpret what you tried to show." (Nathan Yau, "Data Points: Visualization That Means Something", 2013)

"Statistics are meaningless unless they exist in some context. One reason why the indicators have become more central and potent over time is that the longer they have been kept, the easier it is to find useful patterns and points of reference." (Zachary Karabell, "The Leading Indicators: A short history of the numbers that rule our world", 2014)

"The second rule of communication is to know what you want to achieve. Hopefully the aim is to encourage open debate, and informed decision-making. But there seems no harm in repeating yet again that numbers do not speak for themselves; the context, language and graphic design all contribute to the way the communication is received. We have to acknowledge we are telling a story, and it is inevitable that people will make comparisons and judgements, no matter how much we only want to inform and not persuade. All we can do is try to pre-empt inappropriate gut reactions by design or warning." (David Spiegelhalter, "The Art of Statistics: Learning from Data", 2019)

"Agile practices can bring discipline to data science through support for the values and principles of DataOps. However, just as there is no perfect agile framework or set of practices for software development, there is no single set of best agile practices for data science. The right practices to use are context and organization specific and help data analytics teams become more adaptable and collaborative and tighten feedback loops to produce faster (and better) results. The successful application of agile and Lean thinking to data analytics requires observation, constant experimentation, and adjustment." (Harvinder Atwal, "Practical DataOps: Delivering Agile Data Science at Scale", 2020)

"DataOps does not prescribe a particular agile framework, set of practices, artifacts, or roles. Nonetheless, instead of inventing new ways of working, it is better to adapt existing proven practices and combining them as needed. The following are useful agile practices aligned with DataOps principles. However, none are compulsory, and if you find they do not add value, do not use them. Agile frameworks are not mutually exclusive, and practices are situation and context-dependent. You can explore alternatives as long as you stay true to DataOps values and principles." (Harvinder Atwal, "Practical DataOps: Delivering Agile Data Science at Scale", 2020)

"We think of context as the network surrounding a data point of interest that is relevant to a specific AI system. […] AI benefits greatly from context to enable probabilistic decision making for real-time answers, handle adjacent scenarios for broader applicability, and be maximally relevant to a given situation. But all systems, including AI, are only as good as their inputs." (Jesús Barrasa et al, "Knowledge Graphs: Data in Context for Responsive Businesses", 2021)

"A data silo is an isolated source of data that is only accessible to a single line of business (LOB) or department. It leads to inefficiencies, wasted resources, and obstacles in the form of incomplete data profiles and the inability to construct deep insights. [...] On the other hand, a data swamp is a large body of data that is ungoverned and unreliable. It is hard to find data and even harder to use it, which is why it's often used out of context. This is the opposite of data silos in the sense that the data is there and has been brought together, but because it has been done without adequate process and policy, it is as good as not being there. That would be a wasted investment." (Anindita Mahapatra, "Simplifying Data Engineering and Analytics with Delta", 2022

"A well-designed dashboard needs to provide a similar experience; information cannot be placed just anywhere on the dashboard. Charts that relate to one another are usually positioned close to one another. Important charts often appear larger and more visually prominent than less important ones. In other words, there are natural sizes for how a dashboard comprises charts based on the task and context." (Vidya Setlur & Bridget Cogley, "Functional Aesthetics for data visualization", 2022)

"Decision-makers are constantly provided data in the form of numbers or insights, or similar. The challenge is that we tend to believe every number or piece of data we hear, especially when it comes from a trusted source. However, even if the source is trusted and the data is correct, insights from the data are created when we put it in context and apply meaning to it. This means that we may have put incorrect meaning to the data and then made decisions based on that, which is not ideal. This is why anyone involved in the process needs to have the skills to think critically about the data, to try to understand the context, and to understand the complexity of the situation where the answer is not limited to just one specific thing. Critical thinking allows individuals to assess limitations of what was presented, as well as mitigate any cognitive bias that they may have." (Angelika Klidas & Kevin Hanegan, "Data Literacy in Practice", 2022)

"A data product is based on semantically related raw data that is transformed into a meaningful business context and easily discoverable and consumable by business users." (Eberhard Hechler et al, "Data Fabric and Data Mesh Approaches with AI", 2023)

"Data Fabric is a distributed data architecture that connects scattered data across tools and systems with the objective of providing governed access to fit-for-purpose data at speed. Data Fabric focuses on Data Governance, Data Integration, and Self-Service data sharing. It leverages a sophisticated active metadata layer that captures knowledge derived from data and its operations, data relationships, and business context. Data Fabric continuously analyzes data management activities to recommend value-driven improvements. Data Fabric works with both centralized and decentralized data systems and supports diverse operational models." (Sonia Mezzetta, "Principles of Data Fabric: Become a data-driven organization by implementing Data Fabric solutions efficiently", 2023)

"Centralized governance structures often have an abstract view of data, focusing more on uniformity and compliance than context and relevance. While these are essential elements, the nuance often needs to be noticed. Decentralized governance flips the script by giving data ownership to the domain that generates it. The domain has the richest understanding of the data’s context, relevance, and potential impact, thereby being well-positioned to enforce governance policies that improve data quality." (Pradeep Menon, "Data Mesh Principles, patterns, architecture, and strategies for data-driven decision making", 2024)

"Data Mesh addresses data governance challenges by advocating for a federated governance model, which positions accountability for governance with the data owners who are most knowledgeable about the data. In this model, governance is decentralized, with each domain team responsible for the governance of its data products. This approach ensures that governance decisions are made by those who have the deepest understanding of the data’s context, use, and risks. It leads to more relevant, efficient, and effective governance practices that are closely aligned with the specific needs of each domain." (Jean-Georges Perrin & Eric Broda, "Implementing Data Mesh: Principles and Practice to Design, Build, and Implement Data Mesh", 2024)

"When data is considered a product, it creates opportunities for collaboration across different domains. This collaboration involves working with other teams to create, share, and use data products that span multiple areas of expertise, interest, or value. Data Mesh promotes cross-domain collaboration by focusing on the consumers rather than the producers. Data products are made available through standardized interfaces and protocols that support various modes of consumption and are governed by domain experts who understand the context and nuances of their data." (Pradeep Menon, "Data Mesh Principles, patterns, architecture, and strategies for data-driven decision making", 2024)

"At its core, observability rests on three fundamental pillars: metrics, logs, and traces. In the context of data systems, these pillars translate into quantitative measurements (such as data volume, processing latency, and schema changes), detailed event records (including data pipeline execution logs and error messages), and lineage traces that map the flow of data through interconnected processes. Together, they enable a granular and multidimensional understanding of data system behavior, facilitating not just detection but also rapid root-cause analysis." (William Smith, "Soda Core for Modern Data Quality and Observability: The Complete Guide for Developers and Engineers", 2025)

"Data mesh fundamentally reframes data governance and validation by distributing accountability to domain-oriented teams who act as custodians and producers of their respective data products. These teams possess intimate domain knowledge, which is essential for nuanced validation criteria that adapt to the semantics, context, and evolution of their datasets. By treating datasets as first-class products with clear ownership, interfaces, and service-level objectives, data mesh encourages autonomous validation workflows embedded directly within the domains where data originates and is consumed." (William Smith, "Great Expectations for Modern Data Quality: The Complete Guide for Developers and Engineers", 2025)

"Implementing a Medallion architecture isn’t about following a checklist set of instructions. It’s not about taking the exact, fixed process and trying to fit your unique organization around it. Instead, focus on providing your data consumers with context -  organize your data so they understand when it has been cleaned, when it is ready for consumption. Help them find the data to empower their work as fast and easily as possible. Once you have that, you can fit the processes to your design." (Piethein Strengholt, "Building Medallion Architectures: Designing with Delta Lake and Spark", 2025)

18 November 2015

🪙Business Intelligence: Domains (Just the Quotes)

"An important property of a data domain is its scale. The scale determines what relations and operations are possible for the data values in the domain. At the top level, we can differentiate qualitative (or categorical) and quantitative (or numerical) data. At a second level, we can further categorize qualitative data into nominal and ordinal data, and quantitative data into discrete and continuous data." (Christian Tominski & Heidrun Schumann, "Interactive Visual Data Analysis", 2019)

"Data architects often turn to graphs because they are flexible enough to accommodate multiple heterogeneous representations of the same entities as described by each of the source systems. With a graph, it is possible to associate underlying records incrementally as data is discovered. There is no need for big, up-front design, which serves only to hamper business agility. This is important because data fabric integration is not a one-off effort and a graph model remains flexible over the lifetime of the data domains." (Jesús Barrasa et al, "Knowledge Graphs: Data in Context for Responsive Businesses", 2021)

"Data fabrics are general-purpose, organization-wide data access interfaces that offer a connected view of the integrated domains by combining data stored in a local graph with data retrieved on demand from third-party systems. Their job is to provide a sophisticated index and integration points so that they can curate data across silos, offering consistent capabilities regardless of the underlying store (which might or might not be graph based) […]." (Jesús Barrasa et al, "Knowledge Graphs: Data in Context for Responsive Businesses", 2021)

"Data Mesh attempts to strike a balance between team autonomy and inter-term interoperability and collaboration, with a few complementary techniques. It gives domain teams autonomy to have control of their local decision making, such as choosing the best data model for their data products. While it uses the computational governance policies to impose a consistent experience across all data products; for example, standardizing on the data modeling language that all domains utilize." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Data mesh relies on a distributed architecture that consists of domains. Each domain is an independent unit of data and its associated storage and compute components. When an organization contains various product units, each with its own data needs, each product team owns a domain that is operated and governed independently by the product team. […] Data mesh has a unique value proposition, not just offering scale of infrastructure and scenarios but also helping shift the organization’s culture around data," (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022)

"The data mesh pattern doesn't feature a central node and is loosely coupled compared to a hub-spoke architecture. It has different data lakehouse nodes that are independent of each other. The node data lakehouses are domain-driven. A domain can be oriented in multiple ways. The original idea of data mesh alludes to a source-oriented domain aligning to business processes. However, a more practical approach would be to define a domain based on the organizational setup and practicality; for example, a domain can be a product group, it can be separate organizational entities, and it can also be a specific business process, such as marketing. Each domain has its own data lakehouse that is managed and maintained by that domain." (Pradeep Menon, "Data Lakehouse in Action", 2022)

"A domain has two main roles: data product engineer (or just data engineer) and the data product owner (or data product manager, or data steward). These roles can be the same or dedicated people in the domain. Data product owners must have a deep understanding of who their data consumers are, how the data is used, and what methods are used to consume the data. This will help ensure that the data products meet the needs of their use cases. Data product engineers are responsible for creating data products that are high quality, reliable, and usable by consumers. It should be possible to extend existing domain roles to include these domain roles with minimal effort." (Hubert Dulay & Stephen Mooney, "Streaming Data Mesh", 2023)

"To overcome ambiguous domain challenges, each domain boundary must be distinct and explicit. Business area, processes, and data that belong together need to stay together. Additionally, each data domain should belong to one, and only one, Agile or DevOps team. Data integration points within a data domain should be manageable and understood by all team members." (Hubert Dulay & Stephen Mooney, "Streaming Data Mesh", 2023)

"We recommend making domain boundaries concrete and immutable. This helps avoid lengthy discussions about who owns what data, and also prohibits teams from freely interpreting domain boundaries to suit their own needs. Creating a domain-oriented structure is a transition - not only for data, but for people and resources. When creating domain boundaries, resources may eventually align with other teams, disrupting and evolving the current team structure. The entire concept of data mesh is just as much about resource alignment as it is about data, so the realignment of resources should not be considered a roadblock as you go through this process." (Hubert Dulay & Stephen Mooney, "Streaming Data Mesh", 2023)

"When building a data mesh, it is necessary to enable existing engineers in a domain to perform the tasks required. Domains have to capture data from their operational stores, transform (join or enrich, aggregate, balance) that data, and publish their data products to the data mesh. Self-service services are the “easy buttons” necessary to make data mesh easy to adopt with high usability. In summary, the selfservices enable the domain engineers to take on many of the tasks the data engineer was responsible for across all lines of the business. A data mesh not only breaks up the monolithic data lake, but also breaks up the monolithic role of the data engineer into simple tasks the domain engineers can perform." (Hubert Dulay & Stephen Mooney, "Streaming Data Mesh", 2023)

"A data mesh is a decentralized data architecture with four specific characteristics. First, it requires independent teams within designated domains to own their analytical data. Second, in a data mesh, data is treated and served as a product to help the data consumer to discover, trust, and utilize it for whatever purpose they like. Third, it relies on automated infrastructure provisioning. And fourth, it uses governance to ensure that all the independent data products are secure and follow global rules." (James Serra, "Deciphering Data Architectures", 2024)

"At its core, a data fabric is an architectural framework, designed to be employed within one or more domains inside a data mesh. The data mesh, however, is a holistic concept, encompassing technology, strategies, and methodologies." (James Serra, "Deciphering Data Architectures", 2024)

"Data Mesh addresses data governance challenges by advocating for a federated governance model, which positions accountability for governance with the data owners who are most knowledgeable about the data. In this model, governance is decentralized, with each domain team responsible for the governance of its data products. This approach ensures that governance decisions are made by those who have the deepest understanding of the data’s context, use, and risks. It leads to more relevant, efficient, and effective governance practices that are closely aligned with the specific needs of each domain." (Jean-Georges Perrin & Eric Broda, "Implementing Data Mesh: Principles and Practice to Design, Build, and Implement Data Mesh", 2024)

"It is very important to understand that data mesh is a concept, not a technology. It is all about an organizational and cultural shift within companies. The technology used to build a data mesh could follow the modern data warehouse, data fabric, or data lakehouse architecture - or domains could even follow different architectures. (James Serra, "Deciphering Data Architectures", 2024)

"In Fabric, a domain represents a way to logically group data corresponding to specific functional areas. Domains are frequently used to organize data by business sector in order to manage it according to each sector’s regulations, specifics, and requirements." (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

"In the realm of DDD, domains are defined as specific problem spaces that an organization aims to address. These domains encapsulate knowledge, behavior, laws, and activities, and are characterized by semantic coupling - this includes organizational or behavioral dependencies among teams, systems, or services. To simplify management and enhance clarity, domains are often segmented into subdomains, each aligning with different organizational facets." (Piethein Strengholt, "Building Medallion Architectures: Designing with Delta Lake and Spark", 2025)

"[...] the data mesh architecture of Microsoft Fabric primarily supports the organization of data into domains and federated governance [...] Hierarchizing data within OneLake by domain simplifies organizing data, allowing a data producer to easily identify where to deposit data or a data consumer to filter and discover content by functional domain. But it also enables the distribution of governance responsibilities by defining roles and responsibilities for teams in charge of specific domains."  (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)

"The KG serves as the central reference for all structured and unstructured data related to a domain. Because a KG represents information by focusing on the meaning of data, users can overcome challenges related to data types, formats, and provenance, connecting information from multiple data sources. [...] A KG represents the core information and big picture of a domain. Humans can analyze, visualize, and query graph data to extract insights. Inference rules and machine learning algorithms are performed on top of the KG to infer new information not explicitly encoded within the KG. Analysts can use methods such as centrality and connectivity analysis to identify influential nodes, network analysis to detect the shortest path between nodes, and community analysis to recognize groups of similar nodes." (Alessandro Negro et al, "Knowledge Graphs and LLMs in Action", 2026) 

17 November 2015

🪙Business Intelligence: Predictive Analytics (Just the Quotes)

"Put simply, statistics is a range of procedures for gathering, organizing, analyzing and presenting quantitative data. […] Essentially […], statistics is a scientific approach to analyzing numerical data in order to enable us to maximize our interpretation, understanding and use. This means that statistics helps us turn data into information; that is, data that have been interpreted, understood and are useful to the recipient. Put formally, for your project, statistics is the systematic collection and analysis of numerical data, in order to investigate or discover relationships among phenomena so as to explain, predict and control their occurrence." (Reva B Brown & Mark Saunders, "Dealing with Statistics: What You Need to Know", 2008)

"When the purpose of forecasting is to generate accurate forecasts, it is useful to define performance metrics that measure predictive accuracy. Such metrics can tell us how well a particular method performs in general, as well as compared to benchmarks or forecasts from other methods." (Galit Shmueli, "Practical Time Series Forecasting: A Hands-On Guide", 2011)

"Data are essential, but performance improvements and competitive advantage arise from analytics models that allow managers to predict and optimize outcomes. More important, the most effective approach to building a model rarely starts with the data; instead it originates with identifying the business opportunity and determining how the model can improve performance." (Dominic Barton & David Court, "Making Advanced Analytics Work for You", 2012) 

"Most statistical models assume error free measurement, at least of independent (predictor) variables. However, as we all know, measurements are seldom if ever perfect. Particularly when dealing with noisy data such as questionnaire responses or processes which are difficult to measure precisely, we need to pay close attention to the effects of measurement errors. Two characteristics of measurement which are particularly important in psychological measurement are reliability and validity." (Clay Helberg, "Pitfalls of Data Analysis (or How to Avoid Lies and Damned Lies)", 1995)

"Data are essential, but performance improvements and competitive advantage arise from analytics models that allow managers to predict and optimize outcomes. More important, the most effective approach to building a model rarely starts with the data; instead it originates with identifying the business opportunity and determining how the model can improve performance." (Dominic Barton & David Court, "Making Advanced Analytics Work for You", 2012)

"Any presentation of data, whether a simple calculated metric or a complex predictive model, is going to have a set of assumptions and choices that the producer has made to get to the output. The more that these can be made explicit, the more the audience of the data will be open to accepting the message offered by the presenter." (Zach Gemignani et al, "Data Fluency", 2014)

"It is important to remember that predictive data analytics models built using machine learning techniques are tools that we can use to help make better decisions within an organization and are not an end in themselves. It is paramount that, when tasked with creating a predictive model, we fully understand the business problem that this model is being constructed to address and ensure that it does address it." (John D Kelleher et al, "Fundamentals of Machine Learning for Predictive Data Analytics: Algorithms, worked examples, and case studies", 2015)

"Unfortunately, some organizations are replicating the bad data warehouse practice by creating special-purpose data lakes - data lakes to address a specific business need. Resist that urge! Instead, source the data that is needed for that specific business need into an 'analytic sandbox' where the data scientists and the business users can collaborate to find those data variables and analytic models that are better predictors of the business performance. Within the 'analytic sandbox', the organization can bring together (ingest and integrate) the data that it wants to test, build the analytic models, test the model's goodness of fit, acquire new data, refine the analytic models, and retest the goodness of fit." (Billl Schmarzo, "Driving Business Strategies with Data Science: Big Data MBA" 1st Ed., 2015)

"One important thing to bear in mind about the outputs of data science and analytics is that in the vast majority of cases they do not uncover hidden patterns or relationships as if by magic, and in the case of predictive analytics they do not tell us exactly what will happen in the future. Instead, they enable us to forecast what may come. In other words, once we have carried out some modelling there is still a lot of work to do to make sense out of the results obtained, taking into account the constraints and assumptions in the model, as well as considering what an acceptable level of reliability is in each scenario." (Jesús Rogel-Salazar, "Data Science and Analytics with Python", 2017)

"The first myth is that prediction is always based on time-series extrapolation into the future (also known as forecasting). This is not the case: predictive analytics can be applied to generate any type of unknown data, including past and present. In addition, prediction can be applied to non-temporal (time-based) use cases such as disease progression modeling, human relationship modeling, and sentiment analysis for medication adherence, etc. The second myth is that predictive analytics is a guarantor of what will happen in the future. This also is not the case: predictive analytics, due to the nature of the insights they create, are probabilistic and not deterministic. As a result, predictive analytics will not be able to ensure certainty of outcomes." (Prashant Natarajan et al, "Demystifying Big Data and Machine Learning for Healthcare", 2017)

"Data scientists have extensive backgrounds in computer science, coding, machine learning, and statistics. Such an extensive background takes a long time to develop and is usually highly compensated. This has posed a tremendous barrier to many organizations. But this isn’t the only problem with relying solely on data experts. As organizations become more advanced, they are becoming more interested in using data to predict future outcomes. Merely relying on data from the past won’t suffice. Businesses must be forward thinking in how they collect their data to best serve predictive analytics. This means the employees who collect the data (e.g., those rating performance or creating and administering an engagement survey) need to understand how those data will later be analyzed." (Shonna D Watters et al, "The Practical Guide for HR Analytics: Using data to inform, transform, and empower HR decisions", 2019)

"Advanced analytics pivots around machine learning methods. Machine learning employs statistical learning methods to perform analysis on data. These statistical methods utilize algorithms that predict what may happen based on historical data or extract complex mathematical relationships from data to generate insights." (Pradeep Menon, "Data Lakehouse in Action", 2022)

"Many argue that model drift is best monitored by monitoring the data drift in incoming data and the drift in the generated features. As and when the ground truth is available, it is joined by some primary key criteria with the inference data in a Delta table. Again, the update and merge operation support in Delta makes this a breeze. Now the actual and predicted values of the inference data are computed to see how well the model is doing in terms of the quality of insight generation. The feature engineering pipeline is completely in-house and is easier to monitor for drift. The model interpretability may indicate that some columns contributing to the predictive power are incorrect, and it may be necessary to add or remove features. In such cases, a threshold of tolerance is violated, which signals a need for model retraining." (Anindita Mahapatra, "Simplifying Data Engineering and Analytics with Delta", 2022)

"A Data Lakehouse provides a unified platform for various data workloads, such as descriptive, predictive, and prescriptive analytics. It can handle structured and unstructured data and enforce schema at both read and write times, enabling traditional business intelligence tasks and advanced analytics on the same platform." (Pradeep Menon, "Data Mesh Principles, patterns, architecture, and strategies for data-driven decision making", 2024)

"Predictive analytics, involves the use of data, statistical algorithms, and machine learning techniques to identify the likelihood of future outcomes based on historical data. Neural networks, with their remarkable ability to learn and model complex patterns, have become the backbone of modern forecasting methods. Their application ranges from predicting consumer behavior in retail to forecasting the stock market trends, from anticipating weather patterns to foreseeing potential healthcare outbreaks." (Hayden Van Der Post, "Neural Network: Mastering the Art of Algorithmic Trading", 2024)

"Data transformation sits at the heart of every successful data platform, serving as the critical bridge between data ingestion and data consumption. While basic transformations might involve simple cleaning and formatting, advanced transformation techniques encompass complex operations such as data enrichment, sophisticated deduplication, machine learning-based predictions, and the creation of derived metrics that weren’t present in the original data sources. These processes are essential for organizations looking to extract maximum value from their data investments." (Christopher Maneu et al, "The Definitive Guide to Microsoft Fabric From discovery to building a unified, secure, and scalable data platform", 2025)


16 November 2015

🪙Business Intelligence: Data Silos (Just the Quotes)

"Silos are everywhere. […] Silos are common because they are simple, reliable, and unambiguous." (Scott Rosenberg, "Dreaming in Code", 2007)

"Data mart: A subset of a data warehouse that’s usually oriented to a business group or process rather than enterprise-wide views. They have value as part of the overall enterprise data architecture, but can cause problems when they sprout uncontrolled as data silos with their own data definitions, creating data shadow systems." (Rick Sherman, "Business Intelligence Guidebook: From Data Integration to Analytics, 2015)

"Data marts promised to be quicker and cheaper to build, and provided many more benefits - including the benefit of actually being able to finish building them! The data mart was primarily a backlash to the big, cumbersome CDW projects, with the key difference being that its scope was limited to a single business group rather than the entire enterprise. Of course, that shortcut did speed things up, but at the expense of obtaining agreement on consistent data definitions, thereby guaranteeing data silos." (Rick Sherman, "Business Intelligence Guidebook: From Data Integration to Analytics, 2015)

"A data silo is an isolated source of data that is only accessible to a single line of business (LOB) or department. It leads to inefficiencies, wasted resources, and obstacles in the form of incomplete data profiles and the inability to construct deep insights. [...] On the other hand, a data swamp is a large body of data that is ungoverned and unreliable. It is hard to find data and even harder to use it, which is why it's often used out of context. This is the opposite of data silos in the sense that the data is there and has been brought together, but because it has been done without adequate process and policy, it is as good as not being there. That would be a wasted investment." (Anindita Mahapatra, "Simplifying Data Engineering and Analytics with Delta", 2022)

"Since data engineering is such a crucial field, you may be wondering who the main players are and what skill sets they possess. Building a data product involves several folks, all of whom need to come together with seamless handoffs to ensure a successful end product or service is created. It would be a mistake to create silos and increase both the number and complexity of integration points as each additional integration is a potential failure point." (Anindita Mahapatra, "Simplifying Data Engineering and Analytics with Delta", 2022)

"Data silos often start to develop as the gap between data engineering activities and data science activities begins to grow. Data scientists frequently spend the majority of their time creating separate ETL and data pipelines that clean and transform data and prepare it into features for their models. These silos usually develop because the tools and technologies used for data engineering don’t support the same activities for data scientists." (Bennie Haelen & Dan Davis, "Delta Lake: Up and Running - Modern Data Lakehouse Architectures with Delta Lake", 2023)

"Definition of data and AI governance policies, rules, and classifications is critical to break down data silos, allow for a uniform data consumption, and prevent misuse of data. It includes monitoring of compliance and enforcement of data and AI rules and policies on an ongoing basis, as well as ensuring compliance with regulations and laws." (Eberhard Hechler et al, "Data Fabric and Data Mesh Approaches with AI", 2023)

"Where Data Mesh differs from Data Fabric is that it has fixed requirements for the Self-Service platform focused on organizing and managing Data Products by business domain. Another difference is Data Fabric supports managing data as an asset and as a product. A Data Product can be composed of assets that have been governed and managed in a Data Fabric architecture. Data Fabric does not have these fixed requirements, although it inherently supports isolating data and Data Governance enforcement via metadata by business domain. You can think of a Data Mesh Self-Service data platform as supporting separate, independent companies (business domains), although the key criteria are that it does not create data silos and attains data sharing across these companies in a secure, quick, and easy manner. In Data Mesh, Data Products are created and managed by federated business domains and a data platform requires capabilities that enable data and policy federation. This is where a Data Fabric solution can also address Data Mesh’s requirements." (Sonia Mezzetta, "Principles of Data Fabric: Become a data-driven organization by implementing Data Fabric solutions efficiently", 2023)

"Consider data silos. Data silos hinder data accessibility and collaboration, making it difficult to gain a holistic view and leverage the full potential of the available data. They present a real, present, and formidable challenge that almost all data practitioners experience in modern enterprises. Data silos, much like isolated islands in an immense ocean, are repositories of data that are confined within specific departments or systems, disconnected from the broader organizational data landscape. This segregation results in a fragmented data ecosystem, where valuable insights remain untapped, and the collective intelligence of the enterprise is underutilized." (Jean-Georges Perrin & Eric Broda, "Implementing Data Mesh: Principles and Practice to Design, Build, and Implement Data Mesh", 2024)

"Data Mesh advocates for domain-driven ownership of data, enabling individual teams to manage and share their data effectively while aligning with the overall organizational objectives. By embracing this paradigm, enterprises can gradually dismantle the barriers of data silos, paving the way for a more integrated, agile, and data-centric organizational culture." (Jean-Georges Perrin & Eric Broda, "Implementing Data Mesh: Principles and Practice to Design, Build, and Implement Data Mesh", 2024

"Federated computational governance is essential for maintaining consistency and compatibility across the Data Mesh. It ensures that despite the decentralized nature of data ownership, there is a unified framework governing how data is managed, used, and shared. This unified approach is crucial in preventing data silos, ensuring data interoperability, and maintaining the overall integrity of the data ecosystem." (Jean-Georges Perrin & Eric Broda, "Implementing Data Mesh: Principles and Practice to Design, Build, and Implement Data Mesh", 2024)

"Promoting domain-oriented ownership is to combat the common problem of organizational silos. Silos can significantly hinder the free flow of data and expertise, making decision-making and innovation more challenging. We aim to break down these barriers by advocating for domain-oriented ownership and creating a more dynamic and collaborative data management landscape." (Pradeep Menon, "Data Mesh Principles, patterns, architecture, and strategies for data-driven decision making", 2024)

"The ramifications of data silos extend beyond mere inefficiencies; they actively hinder collaboration and innovation within an organization. When data is trapped in silos, it becomes difficult for teams to access the information they need to collaborate effectively. This lack of accessibility and visibility leads to duplicated efforts, inconsistent data practices, and a general sense of organizational disjointedness." (Jean-Georges Perrin & Eric Broda, "Implementing Data Mesh: Principles and Practice to Design, Build, and Implement Data Mesh", 2024)

"[...] organizations are working to unify the data silos that often exist across their data ecosystem. Some are doing this through a centralized physical approach, such as implementing a data lakehouse pattern, which allows them to store and manage all types of data in one place. Many lakehouse implementations such as those from vendors such as Snowflake and Databricks have evolved into what is often referred to as the modern data platform. This is an architectural pattern that combines the lakehouse with tightly integrated tools for data ingestion, transformation, analytics, observability, and governance. This pattern reflects a shift toward cloud-native architectures that are designed to support end-to-end data workflows with scalability and flexibility." (Fern Halper, "Data Makes the World Go 'Round", 2026)

Related Posts Plugin for WordPress, Blogger...

About Me

My photo
Koeln, NRW, Germany
IT Professional with more than 25 years experience in IT in the area of full life-cycle of Web/Desktop/Database Applications Development, Software Engineering, Consultancy, Data Management, Data Quality, Data Migrations, Reporting, ERP implementations & support, Team/Project/IT Management, etc.