Project Outline

[jbusecke]

This should server as a pinned post to keep us organized:

Blogposts

• Datatree
  • Using datatree at all to organize CMIP data (this will basically be the stuff we did at Scipy)
  • Show the rough-and-ready plot from the SciPy lightning talk
  • .plot() method on datatree? (I think this is hard to do in a general way / not a priority)
• Uncertainty
  • Reproduce the plot with matplotlib in full detail (but reduced number of models)
  • Include the error bars, which requires multiple members per model / scenario
  • Requires using datatree in a more advanced way
  • Perhaps also improve the way we are accessing the data (open_datatree vs many open_dataset's)
• Scale
  • Fully reproduce the plot with all available models/members
  • Requires dask to behave
• Visualisation
  • Turn the static plot into an awesome interactive plot
  • Involve holoviews / carbonplan people