Get data on number of buildings in each State

[derekeder]

For the Building electrification part of the state detail page, we will need to get counts of the number of buildings in a given state.

Research where to find this data - bonus points for breaking the building counts up by residential and non-residential.

Data should be collected in a Google Sheet with one row for each State

[derekeder]

Possible data sources

• assessors offices
• CoreLogic
• Open Street Map

[nofurtherinformation]

I took a look at OSM data and chatted with a friend about data sources here. The OSM data wasn't too bad to parse, but the numbers seemed overall quite low. Fortunately, Microsoft released a building footprints dataset that seems more reasonable, and is still OS.

For sanity, here's a screenie of Delaware, one of the places where the two diverge the most:

[nofurtherinformation]

If that data seems reasonable to folks (I think OK on my end), I've got it output in an outer tagged JSON format like the other data so far, and also an OSM script for posterity.

[derekeder]

@nofurtherinformation nice work! The Microsoft data seems like a good find! From the Readme, it looks like they'd like to import it to OSM, but its a bit of a process.

A few follow-up questions:

• is there a way to break the count down by type of building (esp residential vs non-residential)?
• could we sanity-check the data against other sources? should there be a correlation between population?

[nofurtherinformation]

Thanks @derekeder! Hopefully in the future it gets integrated to OSM, but I agree for now Microsoft data seems to be the way.

For sanity checks, the correlation with OSM is already pretty good (~75% correlation, r2 of .56), and comparing against population is also straightforward. There are a bunch of metro counties Assessor data we can compare against (Cook, LA, NY, etc.), and Vermont publishes their State E911 footprint address data, so we can get some regional/rural confirmation too. I suspect Microsoft might over count slightly for things like sheds or accessory units. I'll do a quick sanity check on the counts and circle back here.

For building type, this is slightly trickier. The Microsoft data is just footprint geometries, but we could combine those with zoning data where available to make a decent guess! Places like Texas will be a challenge, since zoning doesn't strictly exist. Same issues exist for OSM coverage on land use/zoning, but we'd get enough buildings that I think it'd be reasonable representative :)

[derekeder]

@nofurtherinformation nice work! Let's roll with the Microsoft data then.

Combining with zoning is a good idea, but as you say, it would vary widely from state to state and as far as i know the data doesn't exist in one combined place like the US Buildings Footprints. It strikes me knowing the number of houses, apartments and other buildings in America is something that a lot of people would want/need to know. I wonder what existing research groups have already done estimates of this.

[nofurtherinformation]

Thanks @derekeder! I've got some time tonight to sanity check the Microsoft data, so it should be good to go for tomorrow evening's meetup.

On building types, I did a bit more digging and found this report from the National Renewable Energy Lab. This goes a bit further down the rabbit hole on modeling (eg. estimating wall types), but they do critically provide by-building metadata files with state tags and building typologies, for commercial these look like:

'SmallOffice', 'RetailStripmall', 'RetailStandalone', 'Warehouse', 'QuickServiceRestaurant', 'Outpatient', 'MediumOffice', 'FullServiceRestaurant', 'SecondarySchool', 'LargeHotel', 'PrimarySchool', 'Hospital', 'SmallHotel', 'LargeOffice'

These are based on 2018 building stocks, so reasonably recent enough. For future reference, here's a link to the ResStock (residential stock) metadata tsv and to the ComStock (commercial stock) metadata tsv. Even if the total building count is different, this is probably a reasonable sample to go on and use as a percentage split on the Microsoft footprints!

[nofurtherinformation]

Even more interesting, the NERL data gives us by building data on things like HVAC system types, energy consumption across a variety of metrics (heating and cooling, lighting) and seasonal estimates. Food for thought--here's a sample row of ComStock data:

bldg_id	105
applicability	True
in.upgrade_name	Baseline
in.tstat_clg_delta_f	5
in.tstat_clg_sp_f	77
in.tstat_htg_delta_f	8
in.tstat_htg_sp_f	63
in.aspect_ratio	2
in.building_subtype
in.county	G0100890
in.building_type	SmallOffice
in.rotation	270
in.number_of_stories	1
in.sqft	17500
in.hvac_system_type	PSZ-AC with electric coil
in.weekday_operating_hours	8.5
in.weekday_opening_time	8
in.weekend_operating_hours	12.5
in.weekend_opening_time	11
in.energy_code_followed_during_last_exterior_lighting_replaceme	ComStock DOE Ref 1980-2004
in.energy_code_followed_during_last_hvac_replacement	ComStock DOE Ref 1980-2004
in.energy_code_followed_during_last_interior_equipment_replacem	ComStock DOE Ref 1980-2004
in.energy_code_followed_during_last_interior_lighting_replaceme	ComStock 90.1-2019
in.energy_code_followed_during_last_roof_replacement	ComStock DOE Ref 1980-2004
in.energy_code_followed_during_last_service_water_heating_repla	ComStock DOE Ref 1980-2004
in.energy_code_followed_during_last_walls_replacement	ComStock DOE Ref 1980-2004
in.energy_code_followed_during_last_windows_replacement	ComStock 90.1-2007
in.energy_code_followed_during_original_building_construction	ComStock DOE Ref 1980-2004
in.heating_fuel	Electricity
in.number_stories	1
in.service_water_heating_fuel	Electricity
stat.air_system_fan_total_efficiency	0
stat.average_boiler_efficiency	0
stat.average_dx_cooling_cop	3.46235472864205
stat.average_dx_heating_cop	0
stat.average_gas_coil_efficiency	0
stat.design_dx_cooling_cop	3.07743026390591
stat.design_dx_heating_cop	0
stat.occupant_density_ppl_per_m_2	0.053819552083549
qoi_report.maximum_daily_timing_shoulder_hour	9.0873786407767
qoi_report.maximum_daily_timing_summer_hour	9.84496124031008
qoi_report.maximum_daily_timing_winter_hour	9.45112781954887
qoi_report.maximum_daily_use_shoulder_kw	23.16081538288
qoi_report.maximum_daily_use_summer_kw	29.7602716030508
qoi_report.maximum_daily_use_winter_kw	29.5302216031377
qoi_report.minimum_daily_use_shoulder_kw	11.7128145227106
qoi_report.minimum_daily_use_summer_kw	11.6088448793368
qoi_report.minimum_daily_use_winter_kw	12.7324554597747
in.nhgis_tract_gisjoin	G0100890003000
in.nhgis_county_gisjoin	G0100890
in.state_name	Alabama
in.state_abbreviation	AL
in.census_division_name	East South Central
in.census_region_name	South
in.weather_file_2018	USA_AL_Huntsville.Madison.723230_2018.epw
in.weather_file_TMY3	Huntsville_Intl_Jones_Field
in.climate_zone_building_america	Mixed-Humid
in.climate_zone_ashrae_2004	3A
in.iso_region	None
in.reeds_balancing_area	89
in.resstock_county_id	AL, Madison County
in.nhgis_puma_gisjoin	G01000302
out.district_cooling.cooling.energy_consumption	0
out.district_cooling.cooling.energy_consumption_intensity	0
out.district_heating.heating.energy_consumption	0
out.district_heating.heating.energy_consumption_intensity	0
out.district_heating.water_systems.energy_consumption	0
out.district_heating.water_systems.energy_consumption_intensity	0
out.electricity.cooling.energy_consumption	7383.33333333333
out.electricity.cooling.energy_consumption_intensity	0.421904761904762
out.electricity.exterior_lighting.energy_consumption	18058.3333333333
out.electricity.exterior_lighting.energy_consumption_intensity	1.03190476190476
out.electricity.fans.energy_consumption	10025
out.electricity.fans.energy_consumption_intensity	0.572857142857143
out.electricity.heat_recovery.energy_consumption	0
out.electricity.heat_recovery.energy_consumption_intensity	0
out.electricity.heat_rejection.energy_consumption	0
out.electricity.heat_rejection.energy_consumption_intensity	0
out.electricity.heating.energy_consumption	7308.33333333333
out.electricity.heating.energy_consumption_intensity	0.417619047619048
out.electricity.interior_equipment.energy_consumption	94127.7777777778
out.electricity.interior_equipment.energy_consumption_intensity	5.37873015873016
out.electricity.interior_lighting.energy_consumption	14847.2222222222
out.electricity.interior_lighting.energy_consumption_intensity	0.848412698412698
out.electricity.pumps.energy_consumption	2.77777777777778
out.electricity.pumps.energy_consumption_intensity	0.00015873015873
out.electricity.refrigeration.energy_consumption	0
out.electricity.refrigeration.energy_consumption_intensity	0
out.electricity.water_systems.energy_consumption	5552.77777777778
out.electricity.water_systems.energy_consumption_intensity	0.317301587301587
out.natural_gas.heating.energy_consumption	0
out.natural_gas.heating.energy_consumption_intensity	0
out.natural_gas.interior_equipment.energy_consumption	0
out.natural_gas.interior_equipment.energy_consumption_intensity	0
out.natural_gas.water_systems.energy_consumption	0
out.natural_gas.water_systems.energy_consumption_intensity	0
out.other_fuel.heating.energy_consumption	0
out.other_fuel.heating.energy_consumption_intensity	0
out.other_fuel.water_systems.energy_consumption	0
out.other_fuel.water_systems.energy_consumption_intensity	0
out.district_cooling.total.energy_consumption	0
out.district_cooling.total.energy_consumption_intensity	0
out.district_heating.total.energy_consumption	0
out.district_heating.total.energy_consumption_intensity	0
out.electricity.total.energy_consumption	157305.555555556
out.electricity.total.energy_consumption_intensity	8.98888888888889
out.site_energy.total.energy_consumption	157305.555463116
out.site_energy.total.energy_consumption_intensity	8.9888888836066
out.natural_gas.total.energy_consumption	0
out.natural_gas.total.energy_consumption_intensity	0
out.other_fuel.total.energy_consumption	0
out.other_fuel.total.energy_consumption_intensity	0
upgrade	0
weight	7.04174071967379
metadata_index	0

[derekeder]

@nofurtherinformation whoa that is some crazy detail! I'd be curious to see what the coverage is for the in.heating_fuel and in.service_water_heating_fuel attributes and what the distribution of values are. if we have good coverage, we could get very precise on how many buildings still need to be electrified!

[nofurtherinformation]

Heya @derekeder, agreed, really detailed! I pulled some summary data by county (we can shift up to state, but may be useful to explore and get a sense of the data). I plopped together a quick explorer with four pages:

• Counts of buildings in each county by residential and commercial
• % of non-electric HCAV/heating (res/com)
• % of non-electric hot water (res/com)
• % of non-electric range/stove (residential only)

Check it out here -- note the color bins on the maps are not fixed between residential and commercial, but this will help to get a sense of the distribution.

For sanity checking data, here's what we've got for counts of buildings via government footprint data vs Microsoft data:

Place	Gov	Microsoft
VERMONT	419,331	351,266
CHICAGO	820,606	907,967
LA	1,122,422	1,33,9971

Taking these 3 cases it's not perfect, but within around 10% or so. I'd be inclined to suggest this is good enough if you feel comfortable, given uncertain on both Microsoft data and open government data, but we can also pull some other county locations to confirm assumptions here.

[derekeder]

@nofurtherinformation thanks for this. This is really great! I agree its close enough that we should proceed with it. do you want to take on getting a CSV of the data rolled up by state with the relevant columns of data?

[nofurtherinformation]

Definitely! Just filed a PR that compiles all this. For convenience, here's a link to the google sheet and a direct CSV link!