Make a cool 'outlier view'

[brifordwylie]

This will be a 'high concept' plot/visualization.

We have a 'smart sample'.. that has 100 samples and then up to 100 outliers in different outlier groups.

Make a scatter plot where we do the following:

For each outlier group (samples will be in the -1 group).

• Compute a median of all the numeric fields for rows in that group
• make a new dataframe that is a 'cluster data' where each row is the median and then a count of the size of the group

[brifordwylie]

We tried several different things for the outlier view, none worked out that well. The current 'smart_sample' gives high/low indications in the 'outlier_group' column. We tried making a visualization of this (scatter plot with sample vs outliers) but nothing seemed that engaging, so throwing this on the backlog and we'll circle back.