Notes from June 2019 workshop

[jorisvandenbossche]

General:

• first morning: mention that people need to download latest version of the material (idea: tag to make a release before the course)
• is %matplotlib inline still needed?

Slides:

• [x] update conda page (environment, source activate -> activate, ..)
• [x] Ipython -> IPython
• [x] IDE's update with Atom / VSCode

Pandas 1:

• [x] column names in overview table titanic wrong
• [x] no .values for dataframe
• [x] html boxen

Pandas 3 selecting:

• [x] and -> & to combine conditions => add tip for this

Pandas 3b indexing:

• [x] update setting with copy warning section with more explanation (temporary variable, boolean selection on population, show both and that the one does not work countries.loc[countries['population'] > 50, 'population'] = 50)

Pandas 4:

• [ ] Ways to shorten this a bit?

Pandas 6 - groupby:

• [x] The color=C0 in the last exercise is not needed anymore

pandas 7 reshape:

• [x] last exercise pivot_table mean -> median

Visualization - plotnine:

• [x] the by in the first pandas example not doing anything?

Visualization - landscape:

• [x] update: Bokeh now supports pdf

Case 1 - bike count:

• [x] have raw data as backup

Case 2 - biodiversity processing:

• [x] groupby doesn't count NaNs, value_counts does -> add a note to explain that difference?

Case 2 - biodiversity analysis:

• [x] don't use built-in sum
• [x] subselection_sex -> make subtask to create this variable

Case 3 bacterial resistance:

• [x] intial tidying: we loose the "experiment id" or "repetion id" in the original data (multiple repeptitions for same phage / genotype, which now is a single row) -> that information is lost
• [x] creation of density_mean -> select column 'optical_density' before .mean() -> no mean of survival etc .. (in solution)

[stijnvanhoey]

• [x] update pandas logo in slides

[stijnvanhoey]

@jorisvandenbossche can we close this issue and maybe move the remaining question to a separate issue?