This is awesome to see! I had a very similar idea recently (representing "big" preprocessing pipelines in Python as DAGs), but for research-scale projects. Fun

This is awesome to see! I had a very similar idea recently (representing "big" preprocessing pipelines in Python as DAGs), but for research-scale projects. Funnily enough, I was also motivated by a time series project that has resulted in several thousand lines of gnarly preprocessing code, and growing - mostly in Pandas. I'm not sure Hamilton would totally work for our use case, but I'll be following closely either way. Thank you for open sourcing this!