Speed up WIT notebook

[BexDunn]

Proposed changes

• Reduced the time period for execution from two years to one year
• Added and documented the ability to have months on the graphs instead of years
• Removed a hard-coded infer_datetime parameter, as pandas doesn't need it anymore and warns if you include it
• added explicit dask chunks

This could probably still go faster, I'm not sure why I'm loading into memory so early here. But this should make a big difference for testing in the meantime

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[robbibt]

Hey @BexDunn, this looks great. For the Dask chunking, I think you can set the x and y chunks to something bigger and it will still have the same effect (Dask will clip the chunks back to the size of your data): e.g. "x": 2048, "y": 2048. That way it will be more generic to other areas/extents in the future (but hopefully still keep the spatial chunking you've set here for this example).

[robbibt]

Looks great!