More Queries + Dashboards

[mrocklin]

We currently run one of the TPCH queries (currently in reduce.py although this location may change), generate a resultset, and store that in a small parquet dataset that is then read by a Streamlit app. This shows both how we can periodically run large dataframe calculations, and serve live results in a flashy way.

Probably we should have a couple more of these, and hopefully some that are both easy to understand and flashy. My guess is that we want to look at the TPC-H queries (these are already pre-formed and solve business problems and we know that they're decently fast) and pick out a couple that we think will look nice and add them.

This will also force us to think about how to have not one computed result, but many, and how to organize them.