Data science projects can be messy, however we don't want to carry that mess over to our production environments.
Data science projects consist of code, data and algorithmic ideas. If these three factors play together there is a tendency that your projects ends up being a total mess.
This article describes csvmodel, a validation tool for small to medium size csv files
A technology perspective
I recently bought a Launchpad Pro Mk3 and noticed that it wasn't all that easy to integrate into my linux setup. Here is how I got it working.
Functional, type save, elegant, ... and it all works in a browser.
Types are really cool
Two different philosophies
Is there really anything like that? Rather some really nice ideas.