Research Task - add route identifiers to shape-aggregated segment speeds

[edasmalchi]

Research Task

• Question or Goal: overarching goal of this issue

The rt_segment_speeds pipeline currently produces two parallel levels of aggregation: by shape and by route/direction

To migrate speedmaps per #1186, I'd like to use the segment speeds data aggregated by shape, but it's also important to have route identifiers (route_short_name, route_id) preserved in that aggregation in order to display route info on the maps and eventual, matching, open data export.

I realize that adding route identifiers to the shape grouping columns could lead to a finer grain of aggregation (generally trips all sharing a route can share multiple shapes, usually at least 2 but there's no guarantee in the GTFS spec that I'm aware of, theoretically trips sharing a shape could share multiple routes).

This seems undesirable. I'd rather propose keeping the grouping columns as-is, and then "aggregate" route identifiers within the groupby (min, max, whatever), keeping the same aggregation grain but with route identifiers that would generally match based on actual practice.

• Expected Outputs / Findings: what do you hope to have created when this is complete?

shape-aggregated segment speeds in pipeline now include route identifiers