Summarize numeric columns

[iaindillingham]

In the dataset-report notebook, @robinyjpark summarizes numeric columns with a CDF (cumulative distribution function). Suggestions from @wjchulme in this Slack thread:

• use a stepped line (plt.step())
• start the curve at 0

[iaindillingham]

@HelenCEBM also suggests summarizing numeric columns including and excluding missing values. This was based on her experience of using cohort-report, which may replace numpy.nan with zeros (check this!). When would a numeric column include a zero that wasn't a measured value?

[robinyjpark]

Update: The dataset-report notebook is now up to date with Will's suggestion (stepped line, start at 0)!