🧭🏭Business Intelligence: Microsoft Fabric (Part VI: Data Stores Comparison)

Business Intelligence Series

Microsoft made available a reference guide for the data stores supported for Microsoft Fabric workloads [1], including the new Fabric SQL database (see previous post). Here's the consolidated table followed by a few aspects to consider: 

Area	Lakehouse	Warehouse	Eventhouse	Fabric SQL database	Power BI Datamart
Data volume	Unlimited	Unlimited	Unlimited	4 TB	Up to 100 GB
Type of data	Unstructured, semi-structured, structured	Structured, semi-structured (JSON)	Unstructured, semi-structured, structured	Structured, semi-structured, unstructured	Structured
Primary developer persona	Data engineer, data scientist	Data warehouse developer, data architect, data engineer, database developer	App developer, data scientist, data engineer	AI developer, App developer, database developer, DB admin	Data scientist, data analyst
Primary dev skill	Spark (Scala, PySpark, Spark SQL, R)	SQL	No code, KQL, SQL	SQL	No code, SQL
Data organized by	Folders and files, databases, and tables	Databases, schemas, and tables	Databases, schemas, and tables	Databases, schemas, tables	Database, tables, queries
Read operations	Spark, T-SQL	T-SQL, Spark*	KQL, T-SQL, Spark	T-SQL	Spark, T-SQL
Write operations	Spark (Scala, PySpark, Spark SQL, R)	T-SQL	KQL, Spark, connector ecosystem	T-SQL	Dataflows, T-SQL
Multi-table transactions	No	Yes	Yes, for multi-table ingestion	Yes, full ACID compliance	No
Primary development interface	Spark notebooks, Spark job definitions	SQL scripts	KQL Queryset, KQL Database	SQL scripts	Power BI
Security	RLS, CLS**, table level (T-SQL), none for Spark	Object level, RLS, CLS, DDL/DML, dynamic data masking	RLS	Object level, RLS, CLS, DDL/DML, dynamic data masking	Built-in RLS editor
Access data via shortcuts	Yes	Yes	Yes	Yes	No
Can be a source for shortcuts	Yes (files and tables)	Yes (tables)	Yes	Yes (tables)	No
Query across items	Yes	Yes	Yes	Yes	No
Advanced analytics	Interface for large-scale data processing, built-in data parallelism, and fault tolerance	Interface for large-scale data processing, built-in data parallelism, and fault tolerance	Time Series native elements, full geo-spatial and query capabilities	T-SQL analytical capabilities, data replicated to delta parquet in OneLake for analytics	Interface for data processing with automated performance tuning
Advanced formatting support	Tables defined using PARQUET, CSV, AVRO, JSON, and any Apache Hive compatible file format	Tables defined using PARQUET, CSV, AVRO, JSON, and any Apache Hive compatible file format	Full indexing for free text and semi-structured data like JSON	Table support for OLTP, JSON, vector, graph, XML, spatial, key-value	Tables defined using PARQUET, CSV, AVRO, JSON, and any Apache Hive compatible file format
Ingestion latency	Available instantly for querying	Available instantly for querying	Queued ingestion, streaming ingestion has a couple of seconds latency	Available instantly for querying	Available instantly for querying

It can be used as a map for what is needed to know for using each feature, respectively to identify how one can use the previous experience, and here I'm referring to the many SQL developers. One must consider also the capabilities and limitations of each storage repository.

However, what I'm missing is some references regarding the performance for data access, especially compared with on-premise workloads. Moreover, the devil hides in details, therefore one must test thoroughly before committing to any of the above choices. For the newest overview please check the referenced documentation!

For lakehouses, the hardest limitation is the lack of multi-table transactions, though that's understandable given its scope. However, probably the most important aspect is whether it can scale with the volume of reads/writes as currently the SQL endpoint seems to lag. 

The warehouse seems to be more versatile, though careful attention needs to be given to its design. 

The Eventhouse opens the door to a wide range of time-based scenarios, though it will be interesting how developers cope with its lack of functionality in some areas. 

Fabric SQL databases are a new addition, and hopefully they'll allow considering a wide range of OLTP scenarios. 

Power BI datamarts have been in preview for a couple of years.

References:
[1] Microsoft Fabric (2024) Microsoft Fabric decision guide: choose a data store [link]

[2] Reitse's blog (2024) Testing Microsoft Fabric Capacity: Data Warehouse vs Lakehouse Performance [link] 

🧭Business Intelligence: Perspectives (Part VIII: Insights - The Complexity Perspective)

Business Intelligence Series

Scientists attempt to discover laws and principles, and for this they conduct experiments, build theories and models rooted in the data they collect. In the business setup, data professionals analyze the data for identifying patterns, trends, outliers or anything else that can lead to new information or knowledge. On one side scientists chose the boundaries of the systems they study, while for data professionals even if the systems are usually given, they can make similar choices. 

In theory, scientists are more flexible in what data they collect, though they might have constraints imposed by the boundaries of their experiments and the tools they use. For data professionals most of the data they need is already there, in the systems the business uses, though the constraints reside in the intrinsic and extrinsic quality of the data, whether the data are fit for the purpose. Both parties need to work around limitations, or attempt to improve the experiments, respectively the systems. 

Even if the data might have different characteristics, this doesn't mean that the methods applied by data professionals can't be used by scientists and vice-versa. The closer data professionals move from Data Analytics to Data Science, the higher the overlap between the business and scientific setup. 

Conversely, the problems data professionals meet have different characteristics. Scientists outlook is directed mainly at the phenomena and processes occurring in nature and society, where randomness, emergence and chaos seem to feel at home. Business processes deal more with predefined controlled structures, cyclicity, higher dependency between processes, feedback and delays. Even if the problems may seem to be different, they can be modeled with systems dynamics. 

Returning to data visualization and the problem of insight, there are multiple questions. Can we use simple designs or characterizations to find the answer to complex problems? Which must be the characteristics of a piece of information or knowledge to generate insight? How can a simple visualization generate an insight moment? 

Appealing to complexity theory, there are several general approaches in handling complexity. One approach resides in answering complexity with complexity. This means building complex data visualizations that attempt to model problem's complexity. For example, this could be done by building a complex model that reflects the problem studied, and build a set of complex visualizations that reflect the different important facets. Many data professionals advise against this approach as it goes against the simplicity principle. On the other hand, starting with something complex and removing the nonessential can prove to be an approachable strategy, even if it involves more effort. 

Another approach resides in reducing the complexity of the problem either by relaxing the constraints, or by breaking the problem into simple problems and addressing each one of them with visualizations. Relaxing the constraints allow studying upon case a more general problem or a linearization of the initial problem. Breaking down the problem into problems that can be easier solved, can help to better understand the general problem though we might lose the sight of emergence and other behavior that characterize complex systems.

Providing simple visualizations to complex problems implies a good understanding of the problem, its solution(s) and the overall context, which frankly is harder to achieve the more complex a problem is. For its understanding a problem requires a minimum of knowledge that needs to be reflected in the visualization(s). Even if some important aspects are assumed as known, they still need to be confirmed by the visualizations, otherwise any deviation from assumptions can lead to a new problem. Therefore, its questionable that simple visualizations can address the complexity of the problems in a general manner. 

Previous Post <<||>> Next Post 

🧭Business Intelligence: Perspectives (Part VI: Data Soup - Reports vs. Data Visualizations)

Business Intelligence Series

Considering visualizations, John Tukey remarked that ‘the greatest value of a picture is when it forces us to notice what we never expected to see’, which is not always the case for many of the graphics and visualizations available in organizations, typically in the form of simple charts and dashboards, quite often with no esthetics or meaning behind.

In general, reports are needed as source for operational activities, in which the details in form of raw or aggregate data are important. As one moves further to the tactical or strategic aspects of a business, visualizations gain in importance especially when they allow encoding data and information, respectively variations, trends or relations in smaller places with minimal loss of information.

There are also different aspects of visualizations that need to be considered. Modern tools allow rapid visualization and interactive navigation of data across different variables which is great as long one knows what is searching for, which is not always the case.

There are junk charts in which the data drowns in graphical elements that bring no value to the reader, in extremis even distorting the message/meaning.

There are graphics/visualizations that attempt bringing together and encoding multiple variables in respect to a theme, and for which a ‘project’ is typically needed as data is not ad-hoc available, don’t have the desired quality or need further transformations to be ready for consumption. Good quality graphics/visualizations require time and a good understanding of the business, which are not necessarily available into the BI/Analytics teams, and unfortunately few organizations do something in that direction, ignoring typically such needs. In this type of environments is stressed the rapid availability of data for decision-making or action-relevant insight, which depends typically on the consumer.

The story-telling capabilities of graphics/visualizations are often exaggerated. Yes, they can tell a story though stories need to be framed into a context/problem, some background and further references need to be provided, while without detailed data the graphics/visualizations are just nice representations in which each consumer understands what he can.

In an ideal world the consumer and the ‘designer’ would work together to identify the important data for the theme considered, to find the appropriate level of detail, respectively the best form of encoding. Such attempts can stop at table-based representations (aka reports), respectively basic or richer forms of graphical representations. One can consider reports as an early stage of the visualization process, with the potential to derive move value when the data allow meaningful graphical representations. Unfortunately, the time, data and knowledge available seldom make this achievable.

In addition, a well-designed report can be used as basis for multiple purposes, while a graphic/visualization can enforce more limitations. Ideal would be when multiple forms of representation (including reports) are combined to harness the value of data. Navigations from visualizations to detailed data can be useful to understand what happens; learning and understanding the various aspects being an iterative process.

It’s also difficult to demonstrate the value of insight derived from visualizations, especially when graphical literacy goes behind the numeracy and statistical literacy - many consumers lacking the skills needed to evaluate numbers and statistics adequately. If for a good artistic movie you need an assistance to enjoy the show and understand the message(s) behind it, the same can be said also about good graphics/visualizations. Moreover, this requires creativity, abstraction-based thinking, and other capabilities to harness the value of representations.

Given the considerable volume of requirements related to the need of basis data, reports will continue to be on high demand in organizations. In exchange visualizations can complement them by providing insights otherwise not available.

Initially published on Medium as answer to a post on Reporting and Visualizations. 

🧭Business Intelligence: Perspectives (Part III: Self-Service BI)

Introduction

According to Gartner, the world's leading information technology research and advisory company, Self-Service BI (aka self-service analytics, ad-hoc analysis, personal analytics), for short SSBI, is a “form of business intelligence (BI) in which line-of-business professionals are enabled and encouraged to perform queries and generate reports on their own, with nominal IT support” [1].

Reading between the lines, SSBI presumes the existence of an infrastructure made of tools to support it (aka self-service BI tools), direct or indirect access to row data and/or data models for the users, and the skillset needed in order to work with data and answer to business problems/questions.

A Little History

The concept of self-service is not new, it just got “rebranded” and transformed into a business opportunity. The need for business users to perform ad-hoc analyses was always there in organizations, especially in the ones not having the right infrastructure for harnessing their data. Even since the 90s with the appearance of products like MS Excel or MS Access in many organizations users were forced by the state of art to learn how to use such products in order to get the answers they needed from the data. Users started building personal solutions, many of them temporary, intended to fill the reporting gaps organizations had. With a little effort and relatively small investment users had the possibility of playing with the data, understanding the data, identifying and solving problems in the business. They acquired thus a certain level of business expertise and data awareness becoming valuable resources in the organization.

With time such solutions grew in scope and data volume, gained broader visibility and reached deeper in organizations, some of them becoming team, departmental or cross-departmental solutions. What grows uncontrolled with time starts to have negative impact on the environment. First tools’ management became a problem because the solutions needed to be backed-up and maintained regularly, then other problems started to surface: security of data, inefficient data processing as increasing volumes of data were processed on local computers and transferred over the network, data and effort were duplicated, different versions of reality existed as different numbers were reported, numbers that were reflecting different definitions, knowledge about the business or data-analysis skillsets. The management needed a more consolidated and standardized effort in order to address these problems. Organizations were forced or embraced the idea of investing money in modern BI solutions, in more powerful servers capable of handling a larger amount of requests, in flexible data models that facilitate data consumption, in data quality initiatives. Thus through various projects a considerable number of such solutions were converted into more standardized and performant BI solutions, the IT department being in control of the changes and new requests.

Back to Present

With IT in control of the reporting requirements the business is forced to rely on the rapidity with which IT is able to address new requirements. Some organizations acquired internal resources in order to build reports and afferent infrastructure in-house, others created partnerships with vendors, or approached a combination of the two. As the volume of requirements isn’t uniform over time, the business has to wait several days between the time a requirement was addressed to IT and a solution was provided. In business terms a few of days of waiting for data can equate with the loss of an opportunity, a decision taken too late, decision that could have broader impact.

A few years ago things started to change when the ad-hoc analysis concept was rebranded as self-service and surfaced as trend. This time vendors like Qlik, Tableau, MicroStrategy or Microsoft, some of the main SSBI vendors, are offering easy to use and rich functionality tools for data integration, visualization and discovery, tools that reflect the advances made in graphics, data storage and processing technologies (e.g. in-memory databases, parallel processing). With just a few drag-and-drops users are able to display details, aggregate data, identify trends and correlations between data. In addition the tools can make use of the existing data models available in data warehouses, data marts and other types of data repositories, including the rich set of open data available on the web.

Looking at the Future

Like its predecessors, SSBI seems to address primarily data analysts and data-aware business users (aka data citizens), however in time is expected to be adopted by more organizations and become more mature where already adopted. Of course, some of the problems from the early days more likely will resurface though through governance, better architectures and tools, integration with other BI capabilities, trainings and awareness most of the problems will be overcome. More likely there will be also organizations in which SSBI will fail. In the end each organization will need to find by itself the value of SSBI.

Previous Post <<||>> Next Post

Resources:
[1] Gartner (2016) Self-Service Analytics [Online] Available from: http://www.gartner.com/it-glossary/self-service-analytics
[2] Gartner (2016) Magic Quadrant for Business Intelligence and Analytics Platforms, by Josh Parenteau, Rita L. Sallam, Cindi Howson, Joao Tapadinhas, Kurt Schlegel, Thomas W. Oestreich [Online] Available from: https://www.gartner.com/doc/reprints?id=1-2XXET8P&ct=160204&st=sb

🧭Business Intelligence: Perspectives (Part II: The Complexity Myth)

Introduction

While looking over “Business Intelligence Concepts and Platform Capabilities” Coursera MOOC resources for Module 2 I run into two similar articles from Solutions Review, respectively Information Age. What caught my attention was the easiness with which the complexity of BI “myth” is approached in both columns.

According to the two sources the capabilities of nowadays BI tools “enabled business users to easily identify and present trends in an impactful way” [1], and “do not require an expert at the helm” [2]. It became thus simpler for users to independently query data and create interactive reports and presentations [2]. In both columns one can read between the lines that the simplicity of using BI tools is equivalent with negating the complexity of BI, which from my point of view is false. In fact here are regarded especially the self-service BI tools, in trend nowadays, that allow users to easily perform ad-hoc analysis with a minimal involvement from IT. Self-service BI is only a subset of what BI for organizations means, and just a capability from the many BI capabilities an organization needs in theory, even if some organizations might use it extensively.

Beyond the Surface

A BI tool is not a BI solution per se, even if many generic BI solutions for different systems are available out of the box. This is one of the biggest confusion managers, users and unfortunately also BI professionals make. A BI tool offers the technological basis for creating a BI infrastructure, though it comes with no guarantees. It takes a well-defined IT and business strategy, one or more successful projects, skillful developers and users in order to harness the BI investment.

On the other side it’s also true that organizations can obtain results also from less, though BI doesn’t equates with any ad-hoc analysis performed by users, even if they use BI tools for this purpose. BI is not only about tools, reporting and revealing trends in the data. BI often implies a holistic knowledge about the business and certain data awareness, without which users will start aggregating and comparing apples with pears and wonder why they taste and look different.

If everything were so simple then why so many BI projects fail to deliver what’s expected? Why so many managers complain that they don’t have the data they need, when they need them? Sure maybe the problem lies in over-complexifying the whole BI landscape and treating everything from a high-level, though that’s more likely not it.

It’s a Teamwork Knowledge Game

BI is or needs to be monitoring and problem solving oriented. This requires a deep understanding about processes and business. There are business users and also BI professionals who don’t have the knowledge one needs in order to approach a business problem. One can see that from the premises they have, the questions they raise, the data they consider, the models they build, and the results.

From a BI professional’s perspective, even if one has a broad knowledge about various businesses, one often lacks the insight in a given business. BI professionals can seldom provide adequate BI solutions without input and feedback from the business. Some BI professionals rely too much on their knowledge, same as the business sometimes expects a maximum output from BI professionals by providing a minimum of input.

Considering the business users, quite often their focus and knowledge cover only the data boundaries of their department, while many problems extend over those boundaries. They know facts that are not necessarily reflected in the data. Even if they are closer to the data than other parties, they still lack some data-awareness (including statistical awareness) in order to approach problems.

Somebody was saying ironically when talking about users’ data and problem solving skills - “not everybody is a Bill Gates or Steve Jobs”. Continuing the idea, one can’t expect users to act as such. For sure there are many business users who are better problem solvers than BI consultants, though on the other side one can’t expect that the average business user will have the same skillset as an experienced BI consultant. This is in fact one of the problems of self-service BI. Probably with time and effort organization will develop such resources, though some help from BI professionals will be still needed. Without a good cooperation between the business and BI professionals an organization might not have the hoped results when investing in BI

More on Complexity

The complexity arises when one tries to make more with the data, especially the data found in raw form. Usually the complexity of raw data can be addressed by building a logical or physical model that allows easier consumption of data. Here is the point where the users find themselves overwhelmed, because for this is required a good knowledge of the physical data model and its semantics, the technical knowledge to build models and the skills to reengineer the logic available in the source systems. These are the themes BI professionals are supposed to excel in. Talking about models, they are the most difficult to build because they reflect various segments of the business, they reflect a breakdown of the complexity. It’s also the point where many BI projects fail as the built models don’t reflect the reality or aren’t capable to answer to business questions.

Coming back to the two columns, I have to point out that the complexity of a subject or domain can’t be judged based on how easy is to approach basic tasks. The complexity lies typically when one goes beyond the basics, when one dives into details. In case of BI its complexity starts when one attempts mixing various technologies and knowledge domains to model and solve daily business problems in an integrated, holistic, aligned, consistent and cost-effective manner. The more the technologies, the knowledge domains and constraints one has to consider, the more complex the BI landscape and solutions become.

On the other side this doesn’t mean that the BI infrastructure can’t be simplified, that BI can’t rely heavily or exclusively on self-service BI solutions. However for each strategy there are advantages and disadvantages and one more likely has to consider both sides of the coin in the process. And self-service BI has its own trade-offs, weaknesses that can be transformed in strengths with time.

Conclusion

When one considers nowadays BI tools capabilities, ad-hoc analyses are relatively easy to perform and can lead to results, though such analyses don’t equate with BI and the simplicity with which they are performed don’t necessarily imply that BI is simple as a whole. When one considers the complexity of nowadays businesses, the more one dives in various problems a business has, the more complex the BI landscape seems. In the end it’s in each organization powers to simplify and harmonize its BI infrastructure to a degree that its business goals aren’t affected negatively.

Previous Post <<||>> Next Post

Resources
[1] Information Age (2015) 5 Myths about Intelligence, by Ben Rossi, [Online] Available from: http://www.information-age.com/technology/information-management/123460271/5-myths-about-business-intelligence 
[2] SolutionsReview (2015) Top 5 Business Intelligence Myths Revealed, by Timothy King, [Online] Available from: http://solutionsreview.com/business-intelligence/top-5-business-intelligence-myths-revealed
[3] Gartner (2016) Magic Quadrant for Business Intelligence and Analytics Platforms, by Josh Parenteau, Rita L. Sallam, Cindi Howson, Joao Tapadinhas, Kurt Schlegel, Thomas W. Oestreich [Online] Available from: https://www.gartner.com/doc/reprints?id=1-2XXET8P&ct=160204&st=sb 
[4] Coursera (2016) Business Intelligence Concepts, Tools, and Applications MOOC, led by Jahangir Karimi, University of Colorado, [Online] Available from: https://www.coursera.org/learn/business-intelligence-tools

🪙Business Intelligence: Data Mesh (Just the quotes)

"Another myth is that we shall have a single source of truth for each concept or entity. […] This is a wonderful idea, and is placed to prevent multiple copies of out-of-date and untrustworthy data. But in reality it’s proved costly, an impediment to scale and speed, or simply unachievable. Data Mesh does not enforce the idea of one source of truth. However, it places multiple practices in place that reduces the likelihood of multiple copies of out-of-date data." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 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 is a solution for organizations that experience scale and complexity, where existing data warehouse or lake solutions have become blockers in their ability to get value from data at scale and across many functions of their business, in a timely fashion and with less friction." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Data Mesh must allow for data models to change continuously without fatal impact to downstream data consumers, or slowing down access to data as a result of synchronizing change of a shared global canonical model. Data Mesh achieves this by localizing change to domains by providing autonomy to domains to model their data based on their most intimate understanding of the business without the need for central coordinations of change to a single shared canonical model." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Data mesh [...] reduces points of centralization that act as coordination bottlenecks. It finds a new way of decomposing the data architecture without slowing the organization down with synchronizations. It removes the gap between where the data originates and where it gets used and removes the accidental complexities - aka pipelines - that happen in between the two planes of data. Data mesh departs from data myths such as a single source of truth, or one tightly controlled canonical data model." (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)

"Data has historically been treated as a second-class citizen, as a form of exhaust or by-product emitted by business applications. This application-first thinking remains the major source of problems in today’s computing environments, leading to ad hoc data pipelines, cobbled together data access mechanisms, and inconsistent sources of similar-yet-different truths. Data mesh addresses these shortcomings head-on, by fundamentally altering the relationships we have with our data. Instead of a secondary by-product, data, and the access to it, is promoted to a first-class citizen on par with any other business service." (Adam Bellemare,"Building an Event-Driven Data Mesh: Patterns for Designing and Building Event-Driven Architectures", 2023)

"Data mesh architectures are inherently decentralized, and significant responsibility is delegated to the data product owners. A data mesh also benefits from a degree of centralization in the form of data product compatibility and common self-service tooling. Differing opinions, preferences, business requirements, legal constraints, technologies, and technical debt are just a few of the many factors that influence how we work together." (Adam Bellemare, "Building an Event-Driven Data Mesh: Patterns for Designing and Building Event-Driven Architectures", 2023)

"The data mesh is an exciting new methodology for managing data at large. The concept foresees an architecture in which data is highly distributed and a future in which scalability is achieved by federating responsibilities. It puts an emphasis on the human factor and addressing the challenges of managing the increasing complexity of data architectures." (Piethein Strengholt, "Data Management at Scale: Modern Data Architecture with Data Mesh and Data Fabric" 2nd Ed., 2023)

"A data mesh splits the boundaries of the exchange of data into multiple data products. This provides a unique opportunity to partially distribute the responsibility of data security. Each data product team can be made responsible for how their data should be accessed and what privacy policies should be applied." (Aniruddha Deswandikar,"Engineering Data Mesh in Azure Cloud", 2024)

"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)

"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)

"To explain a data mesh in one sentence, a data mesh is a centrally managed network of decentralized data products. The data mesh breaks the central data lake into decentralized islands of data that are owned by the teams that generate the data. The data mesh architecture proposes that data be treated like a product, with each team producing its own data/output using its own choice of tools arranged in an architecture that works for them. This team completely owns the data/output they produce and exposes it for others to consume in a way they deem fit for their data." (Aniruddha Deswandikar,"Engineering Data Mesh in Azure Cloud", 2024)

"With all the hype, you would think building a data mesh is the answer to all of these 'problems' with data warehousing. The truth is that while data warehouse projects do fail, it is rarely because they can’t scale enough to handle big data or because the architecture or the technology isn’t capable. Failure is almost always because of problems with the people and/or the process, or that the organization chose the completely wrong technology." (James Serra, "Deciphering Data Architectures", 2024)

📊Business Intelligence: Big Data Analytics (Definitions)

"Big Data Analytics is the process of examining large amounts of data of a variety of types (big data) to uncover hidden patterns, unknown correlations and other useful information using advanced analytic techniques." (Pethuru Raj, "Big Data Analytics Demystified", 2014)

"Big data/analytics is defined as the capability of processing extremely large data sets to identify patterns of relationships (correlation, causality) among data to be used in detecting market trends, consumer behaviour and preferences." (James O Odia & Osaheni T Akpata, "Role of Data Science and Data Analytics in Forensic Accounting and Fraud Detection", 2021)

"Big data analytics is the process of examining large and varied data sets of big data to uncover information including hidden patterns and unknown correlations that can help organizations make better business decisions." (Ahmad M Kabil, Integrating Big Data Technology Into Organizational Decision Support Systems, 2021)

"Big data analytics is the use of advanced techniques to analyze, process and examine big data to uncover hidden patterns, trends and relations in order to assist management decision making." (Steven C S Hui et al, Enhancing Online Repurchase Intention via Application of Big Data Analytics in E-Commerce, 2021)

"Big Data Analytics refers to the intricate process of analyzing vast datasets to uncover hidden patterns, correlations, and customer behaviors from various sources like videos, social networks, and sensors."  (ICT Express, 2023)

"Big data analytics describes the process of uncovering trends, patterns, and correlations in large amounts of raw data to help make data-informed decisions. These processes use familiar statistical analysis techniques - like clustering and regression - and apply them to more extensive datasets with the help of newer tools." (Tableau) [source[

"Big Data Analytics examines large and diverse datasets (i.e. big data) to identify patterns, trends, correlations, and other information that lead to insights organizations can harness in support of better decision-making. Big Data Analytics is the science and engineering of problem solving where the nature, size, and shape of the data renders traditional analytics tools difficult or even impossible to use." (Accenture)

"Big data analytics is the process of evaluating that digital information into useful business intelligence." (Talend) [source]

"Big data analytics refers to the methods, tools, and applications used to collect, process, and derive insights from varied, high-volume, high-velocity data sets." (Microsoft) [source]

"Big data analytics refers to the systematic processing and analysis of large amounts of data and complex data sets, known as big data, to extract valuable insights. Big data analytics allows for the uncovering of trends, patterns and correlations in large amounts of raw data to help analysts make data-informed decisions." (IBM) [source]

🧭Business Intelligence: Enterprise Reporting (Part XI: Between Products, Partners, People and Processes)

Business Intelligence Series

In the previous post, “BI between Potential, Reality, Quality and Stories”, I was commenting five of the important findings of a study led by KPMG in respect to the state of art in BI initiatives. My comments were centered mainly on the first 3 of the 4Ps (Products, People, Partners, respectively Processes) considered in ITSM (IT Service Management). The connection to IT Service Management isn’t accidental, BI being also an organizational capability. Many of the aspects related to the 4Ps perspectives, reveal the maturity of an organization in leveraging its BI infrastructure.  In this post I would like to consider BI landscape from these 4 perspectives.

Products

Products or technology perspective has within BI context a dual nature. First of all we have to consider the BI infrastructure – the whole set of BI tools we have at disposal for our shiny reports. Secondly, because the BI infrastructure doesn’t stand on itself, we have to consider also IT infrastructure on which BI infrastructure is based upon – a full range of ISs (Information Systems) in which data are entered, processed, transported and consumed before they are used by the BI tools. For Data Quality issues, we often have to consider the broader perspective, and tackle the problems at the source. Otherwise we might arrive to treat the symptoms and not the causes. It’s important to note that the two layers or perspectives are interconnected, the consequences being bidirectional.

A typical BI infrastructure revolves around several databases, maybe one or more data warehouses and data marts, and one or more reporting systems. Within the most basic scenario, the data flow is unidirectional from databases to data warehouse/marts, reports being built on top of the data warehouse/marts or directly on the IS’ databases. In more complex scenarios, the data can flow between the various ISs when they were integrated, and even between data warehouses/marts, within a unidirectional or bidirectional flow.  Unless the reports are based directly on the ISs’ databases, such architectures lead to data duplication, conversions between complex schemas, delays between the various layers, to mention just a few of the most important implications. In some point in time the complexity falls down on you.

One of the problems I met is that a considerable percent of the IS are not developed to address BI requirements. It starts with data validation, with the way data are modeled, structured, formatted and made available for BI consumption. If you want to increase the quality of your data, you have sooner or later to address them. It’s important thus the degree to which the systems are designed to cover the BI needs in particular, and decision making in general. This presumes that BI requirements need to be addressed in early phases of implementations, software design or when tools are consider for purchase.

In addition many ISs come with their own (standard) reports or reporting frameworks, becoming thus part of your BI infrastructure, intended or unintended. Even if such reports are intended to cover basic immediate reporting requirements, they not always so easy to consume, the logic behind them is not visible, are hard to extend, are not always tested, the additional reports built in other tools need to be synchronized with them, etc.

Partners

We gather huge volumes of data, we are drowning in it; we want to take decision rooted in data and get visibility into the past, actual and future state of business. How can we achieve that if we don’t have the knowledge and human resources to achieve that? “Partners” is the magic word – external suppliers specialized, in theory, to provide this kind of services: BI analysts and developers, business analysts, data miners, and other IT professionals work together in order to build your BI infrastructure. One detail many people forget is that BI tools provide potentiality; are the skills and knowledge of those working with them that transforms that potentiality into success. On their capabilities depends the success of such projects. Not to forget that BI projects are similar to other IT projects, falling under same type of fallacies plus a few other fallacies of their own derived from exploratory and complex nature of BI projects.

There is a dual nature also in “partners” perspective – except the external perspective which concerns the external partners and the IT department or the business as a whole, there is also the internal perspective in which the IT department plays again a central role. I heard it often loudly affirmed that the other departments are customers of the IT department, or the reciprocal. I have seen also this conception brought to extreme, in which the IT had no word to say in what concerns the IT infrastructure in general, respectively the BI infrastructure in particular. As long the IT department isn’t treated as a business partner, an organization will be more likely sabotaged from inside. Sabotage it’s a word too strong maybe, though it kind of reflects the state of art.

People

Same as partners, people perspective includes a considerable variety of types: IT staff, executives, managers, end-users and other types of stakeholders, each of them with a word to say, grouped in various groups of interests that don’t always converge, situations in which politics plays a major role. It’s actually interesting to see how the decision for a given BI solution is made, how the solution takes its place into the landscape, how it’s used and misused, how personalities and knowledge harness it or stand in the way. I feel that there are organizations (people) which do BI just for the sake of doing something, copying sometimes recipes of success, without uniting the dots, without clear goals and strategy. There are people who juggle with numbers and BI concepts without knowing their meaning and what they involve. This aspect is reflected in how BI tools are selected, implemented and used.

Having the best tools, consultants and highest data quality, won’t guarantee the success of BI initiative without users’ acceptance, without teaching them how to make constructive use of tools and data, on how to use and built models in order to solve the problems the business is confronted with, on how to address strategic, tactical and operational requirements. The transformation from a robot to a knowledge worker doesn’t happen over night. Is needed to make people aware of the various aspects of BI – data quality, process and data ownership, on how models can be used and misused, on how models evolve or become obsolete, how the BI infrastructure has to evolve with the business’ dynamics. There are so many aspects that need to be considered. It’s a continuous learning process.

Processes

In processes' perspective can be depicted a dual nature too. First of all we have to consider the processes which are used to manage efficiently and effectively the whole BI infrastructure. They are widely discussed in various methodologies like ITIL, whose implementation is thoroughly documented. Secondly, it’s the reflection of departmental processes within the various data perspectives – how they are measured, and how the measurements are further used for continuous improvement. 

Considering that this aspect is correlated with an organization’s capability model, I don’t think that many organizations go/rich that far. Sure the trend is to define meaningful KPIs, growth, health and other type of metrics, but the question is – are you using those metrics constructively, are you aligning them with your strategic, tactic and operational goals? I think there is lot of potential in this, though in order to measure processes accordingly is imperative to have also the system designed for this purpose. Back to technological perspective…

 

Previous Post <<||>> Next Post

📉Graphical Representation: Prediction (Just the Quotes)

"Factual science may collect statistics, and make charts. But its predictions are, as has been well said, but past history reversed." (John Dewey, "Art as Experience", 1934)

"The great trouble with all business data upon which the statisticians and economists base their forecasts is that they are ancient history before they ever become available. They pertain to conditions which existed some weeks or months previous. The figures for what is going on at the moment in all lines of business are never available. A business index, while of great interest and value, is always historical and never predictive." (Walter E Weld, "How to Chart; Facts from Figures with Graphs", 1959)

"In part, graphing data needs to be iterative because we often do not know what to expect of the data; a graph can help discover unknown aspects of the data, and once the unknown is known, we frequently find ourselves formulating a new question about the data. Even when we understand the data and are graphing them for presentation, a graph will look different from what we had expected; our mind's eye frequently does not do a good job of predicting what our actual eyes will see." (William S Cleveland, "The Elements of Graphing Data", 1985)

"Changing measures are a particularly common problem with comparisons over time, but measures also can cause problems of their own. [...] We cannot talk about change without making comparisons over time. We cannot avoid such comparisons, nor should we want to. However, there are several basic problems that can affect statistics about change. It is important to consider the problems posed by changing - and sometimes unchanging - measures, and it is also important to recognize the limits of predictions. Claims about change deserve critical inspection; we need to ask ourselves whether apples are being compared to apples - or to very different objects." (Joel Best, "Damned Lies and Statistics: Untangling Numbers from the Media, Politicians, and Activists", 2001)

Statistics can certainly pronounce a fact, but they cannot explain it without an underlying context, or theory. Numbers have an unfortunate tendency to supersede other types of knowing. […] Numbers give the illusion of presenting more truth and precision than they are capable of providing." (Ronald J Baker, "Measure what Matters to Customers: Using Key Predictive Indicators", 2006)

"Data visualization is a means to an end, not an end in itself. It's merely a bridge connecting the messenger to the receiver and its limitations are framed by our own inherent irrationalities, prejudices, assumptions, and irrational tastes. All these factors can undermine the consistency and reliability of any predicted reaction to a given visualization, but that is something we can't realistically influence." (Andy Kirk, "Data Visualization: A successful design process", 2012)

"The term shrinkage is used in regression modeling to denote two ideas. The first meaning relates to the slope of a calibration plot, which is a plot of observed responses against predicted responses. When a dataset is used to fit the model parameters as well as to obtain the calibration plot, the usual estimation process will force the slope of observed versus predicted values to be one. When, however, parameter estimates are derived from one dataset and then applied to predict outcomes on an independent dataset, overfitting will cause the slope of the calibration plot" (i.e., the shrinkage factor ) to be less than one, a result of regression to the mean. Typically, low predictions will be too low and high predictions too high. Predictions near the mean predicted value will usually be quite accurate. The second meaning of shrinkage is a statistical estimation method that preshrinks regression coefficients towards zero so that the calibration plot for new data will not need shrinkage as its calibration slope will be one." (Frank E. Harrell Jr., "Regression Modeling Strategies: With Applications to Linear Models, Logistic and Ordinal Regression, and Survival Analysis" 2nd Ed, 2015)

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)

"Models are formal structures represented in mathematics and diagrams that help us to understand the world. Mastery of models improves your ability to reason, explain, design, communicate, act, predict, and explore." (Scott E Page, "The Model Thinker", 2018)

🧭Business Intelligence: Cognos & SQL Server 2008 - The First Report (Test Drive)

Yesterday evening I installed on my computer the IBM Cognos Express (V9) software, the newest Cognos version, available with a trial period of 30 days. The first attempt to install it was not errorless though as I had to observe several attempts later when I got the same error, the error was not fatal. So, once the instalation complete, the Express Manager started, finding out that I have to install manually the Manager, Advisor, Reporter and Xcelerator - quite easy to do.

Now I was ready to test the tool and create several reports, for this needing to create first a data source - a connection to an existing database with several tables on which the reports will be based. SQL Server 2008's AdventureWorks database seemed to be adequate for this test, in plus I know there are many SQL Server professionals already using it for testing or tutorials.

Step 1: Creating the SQL Server account

Reporting tools need in general only read access to data sources and as a good practice to enforce security and address the special requirements the reports are coming with, so it makes sense to create a new User for this. Open Microsoft SQL Server Management Studio and from the root navigate to Security/New/Login…, this action brining Login – New window in which must be filled

a. the Login name (e.g. cognosuser)
b. the authentification – because Cognos doesn’t supports Windows-based authentication, choose SQL Server authentication, enter the Password and confirm it. By default the ‘Enforce password expiration’ is checked, though this option is not needed for the current requirement so unchek it.
c. Select the Default database, the database you’re intending to base your reports on (e.g. AdventureWorks)
d. Select the ‘Default Language’ (e.g. English)
Navigate to ‘Use Mapping’ tab and in ‘User mapped to this login’ choose the databases you would like to give access (e.g. AdventureWorks) and for each database selected you’ll have to select the schemas in scope (e.g. HumanResources, Person, Production, Purchasing and Sales), while in’Database role membership for’ select the db_datareader database level role name, then you can click ‘Ok’.

 
It’s always a good idea to test the new account before attempting to use it from an external application, this allows catching the eventual issues early in the process, saving your and others’ time!

Step 2: Creating a Data Source

Start the IBM Cognos Express Manager and from Data click on ‘Add…’ button, action that will bring the ‘Enter Data Source Information’ in which you’ll have to select the ‘Data Source Type’, in this case Microsoft SQL Server (SQL 2005 Native Client), fill the ‘Name’ (e.g. AdventureWorks), ‘Server Name’ (here comes your SQL Server name), ‘Database Name’ (e.g. AdventureWorks ), ‘User ID’ (e.g. cognosuser) and password (e.g. something you must guess). As can be seen from the below image, there is also a Databae Object control, which will be populated once you entered the before mentioned information and clicked on ‘Retrieve Database Objects’ button. Select one of the database objects (e.g. AdventureWorks ) and click ‘Ok’, Cognos creating the Data (Source) in the background.

 
Step 3: Creating a Report

Instead of using Report Studio, the standard tool for designing reports, for simple reports or queries could be used instead the Query Studio, an easy-to-use authoring tool, the creation of a report resuming to a simple drag and drop. For this you’ll have to open the Query Studio from the Launch item found on the top menu, this action opening a new browser window, in a next step following to select the package (actually the data source) you want to use. In ‘Insert Data’ mode select a table from the list (e.g. Location) and drag and drop it into the right section, all table’s columns being shown together with a sample set of data.

You could double click on each column and give it a meaningful name – is a good practice to separate the words with spaces though not really necessarily. By clicking on report’s Title you could change its actual value and add a Subtitle if necessary. In ‘Change Layout’ you could change report’s layout by applying different Font, Border and Conditional Styles, much like in Excel, show row numbers by clicking the ‘Show row numbers’ from Change Layout/Set Web Page Size, etc. Save the report once you’ve done the changes, the report will look something like the one from the below image:
Step 4: Running the Report

From ‘Run Report’ Menu item you could choose to run the report mainly in three modes – ‘Run with All Data’, ‘Preview with Limited Data’ or ‘Preview with No Data’. Additionally you can export the report to PDF (Portable Document Format), several versions of Excel (2000, 2002 and 2007), CSV (Comma Separated Values) and XML (eXtensible Markup Language).

Happy coding!

🧭Business Intelligence: Perspectives (Part I: General Issues)

Business Intelligence Series

Introduction

BI projects are noble in intent though many managers and data professionals ignore their implications and prerequisites – data quality (incl. availability), cooperation, maturity, infrastructure, adequate tools and knowledge.

Data Quality

The problem with data starts usually at the source - ERP and other information systems (IS). In theory the system should cover all the basic reporting requirements existing in an enterprise, though that's seldom the case. Therefore, basic reporting needs arrive to be covered by ad-hoc developed tools which often include MS Excel/Access solutions, which are difficult to integrate and manage across organization.

Data Quality (DQ) is maybe the most ignored component in the attempt to build flexible, secure and reliable BI solutions. DQ is based on the validation implemented in source systems and the mechanisms used to cleanse the data before being reported, respectively on the efficiency and effectiveness of existing business processes and best practices.

DQ must be guaranteed for accurate decisions. If the quality is not validated and reviewed periodically, users will be reluctant in using the reports! The reports must be validated as part of the UAT process. Aggregated BI reports need detailed reports that can be used for validation, while the logic and data need to be synchronized accordingly.

The quality of decisions is based on the degree to which data were understood and presented to the decisional factors, though that’s not enough; it's need also a complete perspective, and maybe that’s why some business users prefer to prepare and aggregate data by themselves, the process allowing them in theory to get a deeper understanding of what’s happening.

Cooperation

A BI initiative doesn’t depend only on the effort of a department (usually IT), but on the business as a whole. Unfortunately, the so called partnership is more a theoretical term than a fact, while managers’ and business users' involvement is often suboptimal. 

BI implementations are also dependent on consultants’ skills and the degree to which they understood business’ requirements, on team’s cohesion and other project (management) related prerequisites, respectively on knowledge transfer and training. 

Tools

Most of the BI tools available on the market don’t satisfy all business, respectively users’ requirements. Even if they excel in some features, they lack in others. Usually, more than one BI tool is needed to cover (most of) the requirements. When features are not available, or they are not mature enough, or they are difficult to learn, users will prefer to use tools they already know.

Another important consideration is that BI tools rely on data models, often inflexible from the point of the data they provide, lacking integrating additional datasets, algorithms and customizations. The overall requirements need to be considered more recently from the point of cloud computing technologies, which becomes steadily a requirement for nowadays business dynamics. 

Maturity 

Besides the fact that Capability Maturity Models (CMMs) are difficult to implement, organizations lack the knowledge of transforming data into knowledge, respectively in understanding data and evolving it further in wisdom and competitive advantage. 

Most of the fancy words used by salesmen to sell a product don’t become reality overnight. Of course, a BI tool might have the potentiality of fulfilling the various technical and nontechnical goals, though between a theoretical potentiality and harnessing the respective potential is a long road that need to be addressed at strategical, tactical and operational levels.

Infrastructure

Infrastructure refers to human and technical components and the way they interact in getting the job done. It's not only about "breaking habits" and using the best tools, but in aligning people and technologies to the desired level of performance, of retaining and diffusing knowledge. 

Previous Post <<||>> Next Post

🎯Rukmani Gopalan - Collected Quotes

"A cloud data warehouse is an enterprise data warehouse offered as a managed service (PaaS) on public clouds with optimized integrations for data ingestion, analytics processing, and BI analytics." (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022) 

"Churn refers to rapidly changing the activities and your plan when they are in flux - this is disruptive to your organization and slows your progress. Change refers to an inevitable movement in requirements and helps you plan for and execute this movement thoughtfully." (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022) 

"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)

"If there is one thing I strongly recommend, it is to invest in a cloud data lake and start collecting and processing data that you believe is useful to your organization today." (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022) 

"It’s true that data and data strategy are critical to the organization; however, it’s also true that data by itself is a means to the end of business or customer impact unless you’re a provider of data or data-related services." (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022) 

"Plan for customer impact, and prepare to learn and fine-tune as you progress. Make choices based on the impact they offer to customers, and stay consistent in your implementation while keeping open-minded for learnings. Especially if you are an early adopter of a technology, you can help develop the technology with the provider and thus get ample support from the technology provider in return. Similarly, identify highly motivated early adopters within your customer base and offer to develop your solution with them." (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022) 

"Real-time stream processing refers to the ingestion, processing, and consumption of data with a specific focus on speed, targeting near real time - that is, almost instantaneous results. […] Real-time stream processing pipelines involve data that is arriving from its source at very high velocity; in other words, it is data that is streaming into the system, just like rain or a waterfall." (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022) 

"The lakehouse provides a key advantage over the modern data warehouse by eliminating the need to have two places to store the same data. [...] Data lakehouses offer the key benefit of being able to run performant BI/SQL-based scenarios directly on the data lake, right alongside the other exploratory data science and machine learning scenarios." (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022) 

"The promise of a cloud data lake architecture lies in the boundless diversity of scenarios that it enables." (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022) 

"The very simple definition of cloud data lake storage is a service available as a cloud offering that can serve as a central repository for all kinds of data (structured, unstructured, and semistructured) and can support data and transactions at a large scale." (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022)

"When it comes to data lakes, some things usually stay constant: the storage and processing patterns. Change could come in any of the following ways: Adding new components and processing or consumption patterns to respond to new requirements. […] Optimizing existing architecture for better cost or performance" (Rukmani Gopalan, "The Cloud Data Lake: A Guide to Building Robust Cloud Data Architecture", 2022)

🎯Zhamak Dehghani - Collected Quotes

"A data pipeline is a series of transformation steps (functions) executed as the data flows from one step to another. Data mesh refrains from using pipelines as a top-level architectural paradigm and in between data products. The challenge with pipelines as currently used is that they don’t create clear interfaces, contracts, and abstractions that can be maintained easily as the pipeline complexity complexity grows. Due to lack of abstractions, single failure in the pipeline causes cascading failures." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"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)

"A data product’s primary job is to consume data from upstream sources using its input data ports, transform it, and serve the result as permanently accessible data via its output data ports." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Another myth is that we shall have a single source of truth for each concept or entity. […] This is a wonderful idea, and is placed to prevent multiple copies of out-of-date and untrustworthy data. But in reality it’s proved costly, an impediment to scale and speed, or simply unachievable. Data Mesh does not enforce the idea of one source of truth. However, it places multiple practices in place that reduces the likelihood of multiple copies of out-of-date data." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Data lake architecture suffers from complexity and deterioration. It creates complex and unwieldy pipelines of batch or streaming jobs operated by a central team of hyper-specialized data engineers. It deteriorates over time. Its unmanaged datasets, which are often untrusted and inaccessible, provide little value. The data lineage and dependencies are obscured and hard to track." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Data is a collection of facts put together according to a model. The data model is an approximation of reality, good enough for the (analytical) tasks at hand." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"In addition to limitations of scale, other challenges of data centralization are data quality and resilience to change. This is because business domains and teams that are most familiar with the data are not responsible for data quality." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"[...] the governance function is accountable to define what constitutes data quality and how each data product communicates that in a standard way. It’s no longer accountable for the quality of each data product. The platform team is accountable to build capabilities to validate the quality of the data and communicate its quality metrics, and each domain (data product owner) is accountable to adhere to the quality standards and provide quality data products." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"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)

"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 focuses on the impact of the data and not its volumes. It values data usability, data satisfaction, data availability, and data quality over the volume of the data." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"[...] data mesh introduces a fundamental shift that the owners of the data products must communicate and guarantee an acceptable level of quality and trustworthiness - specific to their domain - as an intrinsic characteristic of their data product. This means cleansing and running automated data integrity tests at the point of the creation of a data product." (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)

"Data mesh is a solution for organizations that experience scale and complexity, where existing data warehouse or lake solutions have become blockers in their ability to get value from data at scale and across many functions of their business, in a timely fashion and with less friction." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Data mesh is an element of a data strategy that fosters a data-driven organization to get value from data at scale." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Data Mesh must allow for data models to change continuously without fatal impact to downstream data consumers, or slowing down access to data as a result of synchronizing change of a shared global canonical model. Data Mesh achieves this by localizing change to domains by providing autonomy to domains to model their data based on their most intimate understanding of the business without the need for central coordinations of change to a single shared canonical model." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Data mesh [...] reduces points of centralization that act as coordination bottlenecks. It finds a new way of decomposing the data architecture without slowing the organization down with synchronizations. It removes the gap between where the data originates and where it gets used and removes the accidental complexities - aka pipelines - that happen in between the two planes of data. Data mesh departs from data myths such as a single source of truth, or one tightly controlled canonical data model." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"In short, a monolithic architecture, technology, and organizational structure are not suitable for analytical data management of large-scale and complex 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)

"One of the limitations of data management solutions today is how we have attempted to manage its unwieldy complexity, how we have decomposed an ever-growing monolithic data platform and team to smaller partitions. We have chosen the path of least resistance, a technical partitioning." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"The distributed nature of data mesh demands immutability to give confidence to data users that (1) there is consistency between multiple data products for a point-in-time piece of data and (2) once they read data at a point in time, that data doesn’t change and they can reliably repeat the reads and processing." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"There are a set of characteristics that can be grouped together as quality. These attributes aren’t intended to define whether a data product is good or bad. They just communicate the threshold of guarantees the data product expects to meet, which may be well within an acceptable range for certain use cases." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Unlike other analytical data management paradigms, data mesh does not embrace the concept of the mythical single source of truth. Every data product provides a truthful portion of the reality - for a particular domain - to the best of its ability, a single slice of truth." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

"Ultimately, Data Mesh’s goal is to enable organizations to thrive in the face of the growth of data sources, growth of data users and use cases, and the increasing change in cadence and complexity. Adopting Data Mesh, organizations must thrive in agility, creating data-driven value while embracing change." (Zhamak Dehghani, "Data Mesh: Delivering Data-Driven Value at Scale", 2021)

🎯Clay Helberg - Collected Quotes

"Another key element in making informative graphs is to avoid confounding design variation with data variation. This means that changes in the scale of the graphic should always correspond to changes in the data being represented." (Clay Helberg, "Pitfalls of Data Analysis (or How to Avoid Lies and Damned Lies)", 1995) 

"Another trouble spot with graphs is multidimensional variation. This occurs where two-dimensional figures are used to represent one-dimensional values. What often happens is that the size of the graphic is scaled both horizontally and vertically according to the value being graphed. However, this results in the area of the graphic varying with the square of the underlying data, causing the eye to read an exaggerated effect in the graph." (Clay Helberg, "Pitfalls of Data Analysis (or How to Avoid Lies and Damned Lies)", 1995) 

"It may be helpful to consider some aspects of statistical thought which might lead many people to be distrustful of it. First of all, statistics requires the ability to consider things from a probabilistic perspective, employing quantitative technical concepts such as 'confidence', 'reliability', 'significance'. This is in contrast to the way non-mathematicians often cast problems: logical, concrete, often dichotomous conceptualizations are the norm: right or wrong, large or small, this or that." (Clay Helberg, "Pitfalls of Data Analysis (or How to Avoid Lies and Damned Lies)", 1995) 

"[...] many non-mathematicians hold quantitative data in a sort of awe. They have been lead to believe that numbers are, or at least should be, unquestionably correct." (Clay Helberg, "Pitfalls of Data Analysis (or How to Avoid Lies and Damned Lies)", 1995) 

"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) 

"Remember that a p-value merely indicates the probability of a particular set of data being generated by the null model - it has little to say about the size of a deviation from that model (especially in the tails of the distribution, where large changes in effect size cause only small changes in p-values)." (Clay Helberg, "Pitfalls of Data Analysis (or How to Avoid Lies and Damned Lies)", 1995)

"There are a number of ways that statistical techniques can be misapplied to problems in the real world. Three of the most common hazards are designing experiments with insufficient power, ignoring measurement error, and performing multiple comparisons." (Clay Helberg, "Pitfalls of Data Analysis (or How to Avoid Lies and Damned Lies)", 1995)

"We can consider three broad classes of statistical pitfalls. The first involves sources of bias. These are conditions or circumstances which affect the external validity of statistical results. The second category is errors in methodology, which can lead to inaccurate or invalid results. The third class of problems concerns interpretation of results, or how statistical results are applied (or misapplied) to real world issues." (Clay Helberg, "Pitfalls of Data Analysis (or How to Avoid Lies and Damned Lies)", 1995) 

References:
[1] Clay Helberg, "Pitfalls of Data Analysis (or How to Avoid Lies and Damned Lies)", 1995 [link]