Isentropic Updates

[dopplershift]

• SkewT solution to 3rd part of exercise doesn't actually call the plot_mixing_lines
• Siphon does import from datetime twice

[dopplershift]

• Isentropic:
  • Hard-coded isobaric--should use metpy accessor
  • Barbs plot has hard-coded PlateCarree rather than data_proj
  • All of the variables should probably be pulled out of the xarray dataset using parse_cf, just for consistency
  • Call to ax.coastlines() and ax.add_feature(cfeature.COASTLINE) in the same cell--should probably nuke all calls to coastlines() from the workshop material to be consistent
  • Labelling contours. I think you can use the return from contour directly to label?

[dopplershift]

Finished in #440