Start Here
Holistics is a business intelligence tool that takes you quickly from data to insights.
The Holistics blog is about the contemporary practice of business intelligence.
We take our commitment to rigor very seriously in our writing. You'll get thoughtful takes on the state of modern data analytics, developments that may change our practice of the field, and historical deep-dives into technologies, philosophies, and stories that may have some influence on the lived experiences of data professionals today.
Read on for a sampling of the blog's best posts:
Data Modeling
One of the core theses of this blog is that the practice of dimensional data modeling needs to change to adapt to the power of modern data stacks. Bill Inmon's Building the Data Warehouse was published in 1990, Ralph Kimball's Data Warehouse Toolkit was published in 1996, and Agile Data Warehouse Design in 2011. All of these books were published before the rise of the powerful, cloud-based, massively-parallel-processing columnar data warehouses. Our belief is that these powerful data warehouses enable vastly different modeling workflows from the past.
We spend a lot of time tracing the history of these ideas, and then canvassing the industry for new adaptations or modifications to them:
- Slowly Changing Dimensions in the Age of the Cloud Data Warehouse
- The Three Types of Fact Tables
- A 2020 Reader's Guide to The Data Warehouse Toolkit
- First Impressions from Agile Data Warehouse Design
The Implications of Massively Parallel Processing (MPP) Columnar Cloud-Based Data Warehouses
A large number of trends in the data world today may be explained by the emergence and broad availability of cloud-based, MPP columnar data warehouses. We explore the implications of this rise on the development of data tools and data practices:
- The Rise and Fall of the OLAP Cube
- Redshift, Snowflake, and the Two Philosophies of Cost
- OLAP != OLAP Cube
- The Data Modeling Layer
- The Two Philosophies of Cost in Data Engineering
The Data Analyst Career
How do you navigate a career in data? How do you better perform in your team? What should you do if you find yourself in a cost-center? How do you evaluate companies before you work for them?
- Data Careers: Dealing with Being in a Cost Center
- Clearer Communication: How To Use The Ladder of Inference When Communicating Data To Your Business Users
- Judge Your Company: Using The Three Levels of Data Analysis For Data Career Decision-Making — figure out if a company is data-driven before you work for them!
- Create a Spreadsheet-Based Data Dictionary
Data Analysts Should Understand How Businesses Work
Data analysts are particularly unique in the tech industry in that they — above and beyond software engineers and designers — must have a base-level understanding of the business in order to function well. We cover the 'bare minimum of business' that data analysts should understand in order to be effective.
- A Comprehensive Summary of Lean Analytics (Part 2)
- When Measuring Performance, Find Lines in the Sand
- Developing an Intuitive Understanding of the SaaS Quick Ratio
- Why Measuring Cash Flow is More Important Than Measuring Profitability
- When Doing Digital Transformation, Think Capabilities, Not Maturity
What Do We Know About Using Metrics?
Business intelligence is fundamentally about using metrics to guide decision-making. There is a remarkably rich literature around the theory and practice of using metrics in business and in society — and so we spend a bit of time writing about these papers and results here:
- The Four Flavors of Goodhart's Law
- Beware What You Measure: The Principle of Pairing Indicators
- The Two Types of Operating Indicators
Standardization Around SQL
It's been clear over the course of the last few years that SQL has emerged as the lingua-franca of modern data analytics. Everything operates with SQL these days. We explore the history of the ideas that led to this trend, and remain watchful for changes to see if this continues to hold true in the future.
DataOps (or Analytical Engineering)
Warning: this is one of the least developed topics in this blog. We're still trying to accurately capture what's going on in the industry here.
There appears to be a slow, barely perceptible shift in the way data analytics is practiced today. There isn't yet a proper name for this collection of best practices — with some people calling it 'dataops' and others calling it 'analytics engineering'. That said, pretty much everyone at the edge of data analytics is observing the same shift. (You may get a better idea of this approach by reading our free guidebook The Analytics Setup Guide — though we are careful to not give it a name there).
We're currently working on unearthing the various elements of this change on the blog: