As mentioned in previous posts, my first big data engineering and data modeling project was for a major health insurer starting in the late 1990’s / early 2000’s. This was the time of the first HIPAA legislation, and the subsequent scramble to comply with that mandate. Much of this project was a data modeling exercise to comply with the HIPAA mandates, but much of it was also an exercise on data modeling for analytics. How to build performant data models that could be used with the Business Intelligence tools at the time, was all greenfield stuff. The Ralph Kimball lifecycle was still in its infancy, and we had tools like Business Objects, Crystal Reports, and Cognos to serve up this data to the end users.
Over the years, there have been many different philosophies on these modeling techniques (Inmon, Bottom-up, Top-down, etc). Each has their own merits, and with some experience you can leverage a hybrid of many of these to accomplish your goals. There has also been an influx of columnar style databases to try and eliminate modeling all together. These come with their own pros and cons too! (incremental updates, row level queries, and transactional workloads don’t work well in these columnar data stores).
There have also been as many different BI tools to facilitate analytics in different ways. Tableau, Looker, Pentaho Business Analytics, Qlik, Power BI, Cognos, Business Objects, etc. The list goes on and on. When I was doing this work 25 years ago I honestly thought everyone would have their analytics figured out within a decade. Man, was I wrong.
There has not been a day gone by when I either speak with a prospect, read an article, or hear nightmare stories about data, and the inability for people to understand it. The problem hasn’t gotten better, it has proliferated. The need to get “actionable insights”, and discern Key Performance Indicators is still a flourishing business. I found this great article here that quantifies many of the problems, as well as some solutions for many Business Analytics projects. It’s worth the short read!
Here at kpi-forge we are currently working on some smaller projects that fit very well into the old school Kimball/RDBMS/ETL arena. We’re solving business problems quickly, efficiently, and making the venerable “lightbulb go off” when delivering these KPI’s to our customers in record time. There are many problems to be solved, and we’re ready to help solve them for you! Let’s talk today!
Oliver Nielsen
CEO/Founder
KPI Forge
Comments