"Searching multivariate data with BayesDB" tutorial notebook

[zane]

The Juypter notebook demo of data search functionality on COMPAS. This notebook demo, to be prepared by @Sarava-007 with help from her UROPs, will show:

1. loading the data and guessing types and building models and looking at a dependence heatmap as a diagnostic of whether the models are totally crazy or not ("Search" notebook should show loading data into BayesDB #8, "Search" notebook supports building an ensemble of models #9, "Search" notebook supports using a dependence heatmap as a diagnostic tool #10)
2. filtering convicts via SQL operators rapidly leads to very few matches ("Search" notebook should show inadequacy of SQL #11)
3. using combinations of SQL and BQL-similarity-with-respect-to leads to sensible matches. ("Search" notebook should show how using SQL and BQL in combination improves results #12)
4. it is possible to search by similarity to hypothetical rows. Warning: @zane, this might require @fsaad to add a new feature, or debug/refactor an existing feature. ("Search" notebook should demonstrate search by similarity to a hypothetical row #13)
5. it is possible to search for anomalous/unlikely records ("Search" notebook should demonstrate finding unlikely rows #14)

[zane]

Migrated this information to a milestone instead.