Airbnb/Rental Supply Data Collaborative
Project Sponsor: Charlottesville HAC (via Chris Meyer, Heather Hill, Sally Hudson)
Project team: Nathan Day, Matt Thomas, Courtney Whalen
The below report was submitted to HAC for its meeting packet in March (just as COVID-19 shut downs were beginning). That meeting did not happen but may this July. Not submitted to the HAC but created by the collaborative is an interactive visualization of short term rental properties based on public data. There is some offset on the timing of these datasets. The business licenses are those on file at the end of 2019. The zoning data represents all the homestays approved this year (2020). There are most likely some new property owners doing homestays this year that the Commissioner of Revenye did not have in 2019. Additionally, the homestay permit regulates short term rentals in owner occupied properties, primarily in the R-1 and R-2 zoning districts. Short term rentals are permitted by right in every zoning district which allows for hotels (so no zoning application is needed). The following zoning districts permit short term rentals as a by-right use: D, DE, WME, WEW, CH, HS, HW, WSD, URB, SS CD, CC, B-2, B-3.
You can review our visualization here: https://smartcville.com/airbnb/
Data Collaborative Report
View this report as a Google Doc or save as PDF
Question posed by HAC:
To what extent are properties in Charlottesville being used as full-time Airbnb properties?
Process and Timeline:
November 2019
– CCI invited to project by HAC member Chris Meyer
December 2019
– CCI establishes data collaborative that includes a resident data scientist and UVA School of Data Science student
– Launch meeting occurs to learn more about the question to investigate. Attendees include the initial team and one resident, the Zoning Administrator, and one Planning Commissioner.
– A request is made to the Zoning Administrator and Commissioner of Revenue for public data related to the project
– Another resident joins the team to provide technical support
– Data collaborative members begin holding regular meetings to develop the project
January 2020
– Technical team shares progress Chris Meyer and one resident
February 2020
– Data is received from the Zoning Administrator and Commissioner of Revenue
– Technical lead shares progress with Chris Meyer
March 2020
– Charts are sent to Zoning Administrator and Commissioner of Revenue for feedback
– Project manager meets with Chris Meyer to discuss presentation at HAC meeting
Methodology:
Airbnb does not provide the data requested by HAC. The team used a method called “scraping” to extract data about City-based Airbnb properties. This is a notoriously difficult thing to do. Airbnb and VRBO both make scraping data from their site challenging. In addition, geolocation data scraped from the sites is close but imprecise.
That said, the team was able to set-up scraping methods to extract information about the total number of properties in the City (FIGURE B). We believe this is a “floor” total that does not include properties that are temporarily inactive (seasonal or de minimus properties).
The team was also able to set-up scraping methods to determine properties offering full weekday listings (Monday-Friday). We scraped these listings for three consecutive months. Each week was intentionally chosen to avoid times where a weekday stay may be popular (like graduation or TomTom). The weeks chosen were the first full calendar weeks in October, November, December. The results can be seen in Figure C. We believe this is a “floor” total because, during our informal exploration, we found evidence both that some listings had not posted availability that far out and that suspected bookings for some may have occurred.
We’ve also used public data to place each Airbnb listing into by-right and restricted zones
We feel strongly that the scraped dataset has provided sufficient data to use for a broad analysis of the scope of this type of short term rental usage. Because of the challenges in scraping this data (non-exact location information) the zoning breakdowns are not exact counts, but we believe the ratios are durable and good approximations of the situation in Charlottesville.
Discussion and Next Steps:
This data is simply a first step for the HAC. It does not represent anywhere near a full analysis. However, when combined with the institutional knowledge of the HAC, expertise of City staff, and the qualitative context of residents it can be used to inform next steps or potential future questions.
FIGURE A
Baseline numbers – single family homes in City housing stock*
| Type | Count |
| Single Family Detached | 8308 |
| Single Family Detached w/1 ADU | 863 |
| Single Family Detached w/2 ADUs | 67 |
| Single Family Detached w/3 ADUs | 25 |
*Source: City of Charlottesville
FIGURE B
FIGURE C
