Possible fix for #356

[cisaacstern]

closes #356

This works, but there may be a better way to do this, thus "possible" in the title.

And specifically, perhaps @atmorling has some insight into any relationship between #356 and the fix for #352 (which may point to a better solution).

[atmorling]

I think something like this makes sense:

events_df["time"] = events_df["geojson"].apply(lambda x: ptime(x.get("properties").get("datetime"))
events_df = clean_time_cols(events_df)

Basically tease out the time data from the geojson before we do clean_time_cols
Then we shouldn't need any extra formatting in this function.

[cisaacstern]

Oh cool, yeah if that works that sounds better! I figured you'd have a good insight on this.

[cisaacstern]

If you'd like me to make a stab at that I can try, though tbh if you're keen, I think you're in a better place in re: familiarity with this section of code. So I'd be happy to just close this PR and have it superseded by your proposed fix.

[atmorling]

Sure, will do!

[atmorling]

Superseded by #358