The good news is that I’m spending some time learning matplotlib, numpy, and Cassandra.
The bad news is that I’m not working on Ancho. Instead, I’m working on a project for Custom Insurance Services. You see, we switched to a new agency management system, which like all major software packages, has its own particular problems. Among other things, its built-in reporting tools are pretty awful. They are both uninformative and unreadable.
Fortunately, it can spit out most of the raw data needed to produce the reports I actually want. So, I’m writing a tool that slurps up the raw data, normalizes it, generates some rollups, and uses those rollups and Sphinx to generate reports in EPUB format with embedded matplotlib charts.
Is that the best way to do it? Who knows? I doubt it, but it’s the way that allows me to spend time learning the other technologies that I need to know how to use.