Data Engineering

Haven't heard of this before, but I've basically implemented it in my own work and it's worked out great.

I believe implementing facts and dimensions via a star schema is still very useful and serves as a great modeling format, however there's definitely times where my downstream users need a single OBT type table. Ex. CSMs use a tool fed by data warehouse but none of them know SQL, so we feed a single "bridges" table to them that includes everything.

One interesting flaw with my implementation I've found, my users are looking for all the data under the sun, and I wasn't previously prepared originally and had to add a lot of columns over time. This grew into a massive mess that needs to be remodeled via multiple OBT type bridge tables. It's very easy to just throw everything in these tables and lose all structure.