A new picture of the SF housing crisis using zone and building heights
By: Asheem Mamoowala
San Francisco has been in the midst of a housing shortage since the early 1990’s. While there isn’t much room to expand horizontally, I was curious how much room there was to grow vertically. That led me to ask the question “Given the existing zoning rules, are SF buildings as tall as they could be?”
I couldn’t find any maps that combined the zoned height information with the as-built height information. San Francisco’s open data portal, DataSF, provides datasets for Height and Bulk Districts and Building Footprints. So for a recent internal hack day, I decided to merge this data and visualize the difference between allowed and actual building heights.
This map is a naive visualization of buildings that have room to add at least one floor while remaining within their zoned height restrictions. I used a height of 2.75m (9ft) to represent one floor, which isn’t applicable to all types of buildings but gives us a broad metric to apply. There are many laws and regulations beyond just zoned height that determine the final building footprint — this map shows us simply where there’s room to grow.
What immediately stood out to me is that most of the residential neighborhoods have the capacity to add an additional floor to their buildings without any change to zoning policy. Currently, city ordinance 162–16 forbids adding additional dwelling units (ADUs) through building height increases. It’s a constant struggle for cities, especially SF, to manage infill development without changing the character of the city.
Pan through the map below to see unused height allowances across SF, what stands out to you?
How I built it
Initial Prototype
I started by downloading the Bulk Height districts as a GeoJSON file and making an interactive prototype to see if there was, in fact, a difference between built height and allowed height of existing buildings.
To do this, I used a mousemove event handler along with map.queryRenderedFeatures() to compute the join whenever the user hovered over a building.
Merging datasets
The DataSF datasets are served using the Socrata Open Data API (SODA) which allows geospatial queries including polygon intersects. This allowed creating a geospatial join tool to generate the intersection of bulk height zones with building footprints and merging the relevant properties.
The merged data was written back out as a GeoJSON file with only the required properties. Using Studio, I was able to upload this file and convert it into a vector tileset which loads and renders faster than using the file directly as a GeoJSON source.
Styling the Map
I used data-driven styling with multiple fill extrusion layers to display the existing building height and the unused height allowance.
- actual-height layer uses a filter to select only buildings that have a height_delta greater that one floor. This layer shows the actual height of existing buildings
- allowed-height layer uses the same filter, with a stop function to create a color gradient for the additional allowed height for buildings. The fill-extrusion-base property is set to the building’s existing height, to create the 3D overlay.
Explore the building heights map for yourself. Interactive maps with open data from cities allow us to understand better our urban environments and the policies that shape them. Check out our tools for data visualization and how we’re helping global cities address their most pressing needs.
Are San Francisco buildings as tall as they could be? was originally published in Points of interest on Medium, where people are continuing the conversation by highlighting and responding to this story.