Handling missing data in linear regression on timeseries data

[AlexAndorra]

Hi,
Maybe casting your data to a numpy masked array could help? That way, PyMC should be able to handle and impute the missing data points. Here is an example of how to do that.
Hope this helps