Using Pandas within Pymc3 context?

[chartl]

Possibly. Try running with a subset .values[:10000,:] just to see if it works. If it does, the issue may be due to the data size.