Document commonly used additional bldg metadata columns

[jpvelez]

What

An evaluation and comparison of the supplementary bldg_id columns that ResStock users routinely create.

The initial scope of this ticket is:

• "building type" columns derived from in.geometry_building_type_acs
• "heating setup" columns derived from in.hvac_heating_type (technology) and hvac_heating_fuel
• And string to numeric conversions, like occupants or income

In addition to the above, RMI also consolidates:

• in.vintage_acs
• in.water_heater_efficiency
• in.infiltration
• in.hvac_heating_efficiency
• in.hvac_cooling_efficiency
• in.hvac_cooling_type

Not in initial scope:

• @brmurray's upgrade description columns, as there are release-specific. Also, these may be superceded by using some of RMI's columns above.

Why

RMI and Switchbox both routinely add additional bldg_id columns that they use to slice and dice results. It would to:

1. Move this to a dataset that is fetchable with bsf, rather than applying this logic every time
2. Learn from each other about why we do these
3. Consolidate where possible

How

• Look at @brmurray R code in reports/lib
• Look at Al's python code in comparisons repo
• Document additional columns created by each team, their unique values, and how they are constructed out of existing columns
• Also document why each team aggregates things in this way: varying definitions of multifamily, etc.

Deliverables

• Comments in this issue