featurizer × chicago311
Which 311 requests will resolve slowly? (requests ← area/type demand streams)
1 · Full-cohort materialization (v1.0.0)
Each variant materialized through to_dataframe(connection=…)
at one as-of date (2020-01-01 — the day
after the last knowledge date, so every event is under the window); no
permanent writes. Variant definitions and the recovered wide transformer set:
benchmarks/final_matrix.py.
| variant | features | shards | rows | dup-names | seconds | was (v0.8.0) | status |
|---|---|---|---|---|---|---|---|
| narrow | 28 | 1 | 30,654 | 0 | 0.8 | 0.8 | materialized |
| all-agg | 191 | 1 | 30,654 | 0 | 5.7 | 6.0 | materialized |
| wide | 907 | 1 | 30,654 | 0 | 47.5 | 49.2 | materialized |