At work we need to check loans to see if they qualify for CRA. Right now I am in the process of getting us off of the terrible CRA Wiz product (apparently, I have no personal experience but hear lots of griping).
Steps in this project
- Installing PostGIS and enabling in our existing Postgres db.
- Downloading and install the state, county, address, and census lines
- Creating some bulk processing db sprocs
- Ruby on Rails project to take inputs/models and feed them through a data clarification pipeline
Lessons learned
- Always get the bigger HD. The data is over 100GB once unpacked into PG. And you need swap memory. Even worse is my VM is snapshotting deltas and consuming ~2x the disk needed. I really wish I had this project on Docker. I wasted like 8+ hours looking for swap space and cleaning up my HD.
- Census.gov/TIGER does a pretty cool service by open-sourcing this data. In fact, I’m fairly certain you could build a product on just slapping together a decent web-api with performance standards and documentation on top of this DB. (SmartySheets?)
- Ruby is just awesome. Falling even more in love with Ruby after this project. It is just so damn elegant. There are some quirks I don’t love about rails, but nothing is perfect.
