Idea for using more Seaborn and less Pandas in Python Data Vis

[clairecahoon]

For the Data Visualization with Python workshop:
We could take out most of the example plots from the pandas section and move them to the seaborn section, then add an explanation of the differences in between the sections (including that most of the plots could be done in either library). The idea is that you're just showing that it's possible to use Pandas and Pandas features for exploration, and then focusing on a library that has more options for a polished design. I don't know if all of it's possible, but here's a sample outline:

1. Datasets and Pandas
2. Visualization
3. Univariate plotting with Pandas
   a. Bar and line plots
4. Bivariate plotting with Pandas
   a. Scatter plots
   b. Stacked plots (with pivot tables)
5. Styling with Pandas
6. Seaborn
   a. When you might use Seaborn vs. Pandas
7. Univariate plotting with Seaborn
   a. Histograms
   b. Boxplots
   c. Line charts
   d. Time series
   e. Area charts
8. Bivariate plotting with Seaborn
   a. Boxplots (again)
   b. Hexplots
   c. Hexplots with accompanying histograms
   d. Faceting
9. Saving a Figure
10. A brief look at Bokeh
11. Other Libraries