The Housing Insights database aims to capture a complete inventory of all the subsidized affordable housing available in DC. The data sources that we are using include:
1) The Preservation Catalog
2) DC’s Open Data data set of affordable housing properties
3) The Quickbase Database maintained by the Department of Housing and Community Development (DHCD)
In the future, the following additional data sources might also be useful to include:
4) The Housing Production Trust Fund annual report includes information such as building name, address, funding level, total project funding amount, etc. Recent buildings (definitely post 2015 and mostly 2011 and later) are available from DHCD’s quickbase API. It may be worthwhile to scrape the PDF versions of HPTF annual reports for earlier years if needed, and ideally, to cross-reference the PDF data with DHCD’s database in order to look for inconsistencies and add additional data as needed.
5) PUD history from the office of zoning - the Affordable Housing Data Set from opendata.dc.gov reportedly includes most inclusionary zoning units. It is only missing those that are part of the Affordable Dwelling Units program, because ADU units typically went through the PUD process outside of the inclusionary zoning program (according to the Office of the Deputy Mayor for Planning and Economic Development). That said, individual Planned Unit Development applications often include details about the quantity and type of affordable units being included in each PUD. Further analysis would be required in order to determine whether affordable housing details extracted from PUD applications might contain any additional relevant data that is not already included in the Affordable Housing Dataset.
6) Although the Affordable Housing Data Set on opendata.dc.gov reflects most inclusionary zoning units from 2015 to the present, the Deputy Mayor for Planning and Economic Development also has historical data on inclusionary zoning projects prior to 2015. DMPED previously indicated it might be possible to obtain a one-time pull of that historical data.
7) DHCD may have additional historical data on older buildings not currently included in their Quickbase database (which only includes data starting in October 2010)
8) It may be possible to obtain additional data via direct request to the DC Housing Finance Agency (DCHFA) as well as to the DC Housing Authority (DCHA)
Compiles all the federal programs, with sporadic additions of older DC-funded buildings
DMPED manages a Quickbase database that pulls from the DMPED, DHCD, DCHFA and DCHA data (some from APIs, and some from manual sources). They’ve put that data from the affordable housing database onto opendata.dc.gov
Some notes on the Affordable Housing data source opendata.dc.gov dataset scope:
This database is connected to DHCD’s internal record keeping; however, per DHCD’s admission it is an incomplete account of buildings they have funded. Per DHCD it should contain all post 2015 buildings, about “80% of buildings they worked on 2013-2015, 50% of those from 2011-2013, and sporadic before that.”
Although the majority of the buildings in the Housing Insights database are derived from the Preservation Catalog, the data that we have ingested from opendata.dc.gov as well as from the DHCD database have enabled us to incorporate additional buildings that were not represented in PresCat originally. Many of these additional buildings were identified via records from local (vice federal) subsidy programs, since PresCat is primarily based on federal subsidy data.
In PresCat, every building has a unique ID number, however since buildings that we have identified via DHCD or other sources may not have been in PresCat previously, those buildings lack a PresCat ID number. Since we want to make sure that each building in our database gets its own unique ID number (so that we can consolidate all the details for any given building in a single record), we need a way to match any additional buildings that we discover from new sources against the buildings already represented in our database. If the building matches one already in our database, we add whatever new details we have to the existing record. If no match is found, then we create a new record (and a new ID) for that building in our database.
We use DC’s Master Address Repository to determine whether a building from a new datasource is a match to a building that already exists in our database, or whether we need to create a new record for it. The MAR enables us to check for matches based on building address, parcel square/suffix/lot (SSL), intersection, block or place name.