Building an affordable housing data portal with the City of Austin
In the spring of 2018, we were awarded a 6-month fellowship with Code for America to work on the affordable housing crisis with the City of Austin. Our primary focus was to figure out what technology Austin needed most to help the situation.
Background
If you’d like more context and background about that fellowship and our work during it, check out the following articles and video of our presentation at the Code for America Summit conference:
- Our plan to tackle affordable housing in Austin, TX
- Building an affordable housing application starts with data
The Problem
There are quite a few issues in the affordable housing space, but the one that we chose to focus on with our fellowship was the lack of infrastructure for storing/sharing affordable housing data. It was a fairly straightforward problem whose solution could lay a foundation for identifying and solving more complex issues.
Between non-profits and government, there are a lot of affordable housing agencies in Austin. However there wasn’t a unified method of tracking affordable housing data across all those agencies. Each organization had their own method of tracking information (if there was one at all). It was a classic story of data living in silos, which resulted in missing information and duplicated work.
These data silos didn’t just cause issues with efficiency, they affected how much data the government had for decision making. Affordable housing is a very large and complex problem so quality data can be incredibly powerful. Building new housing is also a slow moving process, which makes good decision making that much more important. Better data helps support the decision making process by allowing the government to understand where housing currently is, where residents want to be, and where it will be in the years to come.
Our Solution
The most important part of any project involving multiple stakeholders is getting everyone in a room together to come to a consensus. Fortunately, the housing agencies in the City of Austin were already meeting and discussing these issues and potential solutions before we came on board. Our fellowship provided the extra resources needed to move the project forward.
We took all the research and conclusions made by the housing agencies and helped to figure out the next steps in building a solution. We worked closely with a few stakeholders who had a good understanding of what all the other players needed. After a few iterations we landed on the concept to build an affordable housing data hub for the City of Austin. This data hub would store the housing data for all agencies in the city and would be accessible via an API so that any other applications could use the data as needed.
Step 1: Gather all the data
The first step to making this project happen was to gather and combine all the existing datasets from the different housing agencies. This was no easy feat. Each dataset had its own format, fields, and had varying levels of accuracy and completeness.
We then had to figure out which fields would ultimately end up in the data hub. We worked with our stakeholders to determine a final list of fields and how they mapped to the fields in the existing datasets.
Once we knew what the fields would be, we had to combine all the existing datasets into one big dataset and remove any duplicate properties. We finally had a full set of data that could be loaded into the database.
Step 2: Build the API
Once all the data had been loaded into the database, we needed to build an API to access the data. We chose to build an API to ensure that the data hub could easily synchronize with multiple systems and wouldn’t be attached to one specific application moving forward. The API was built using a RESTful architecture.
Step 3: Build the data portal
Once the API was ready to go, we focused on the affordable housing data portal. The data portal is a web application that connects to the data hub’s API and allows housing agencies to view and maintain the data.
Upon logging in, the user is presented with a list of basic information for all properties in the portal. Search and filter tools allow them to find a property quickly. New properties can also be added from this page.
Each property has much more information than what is displayed on the listing page. The property detail view displays all the fields for the property and allows the user to edit any of the fields.
When editing a property, each field is verified to have the correct format and type of data. Information icons next to each field give more details as to what type of data should be entered. This helps ensure that the data in the hub is as consistent as possible.
Fields that are empty or out-of-date are highlighted in red. The system stores the date a field is last updated, automatically calculates when it next needs to be verified, and alerts users when it goes out of date. Each field ‘expires’ at a different rate depending on how likely it is to change. For example, the street address will never expire but whether or not pets are allowed will expire periodically.
Users can also add any number of notes to each field, allowing for more granular or exceptional details about the field to be stored for reference.
The other major piece of this data hub is a reporting tool that will allow housing agencies to better understand the supply of affordable housing in Austin. This tool contains multiple data visualizations and custom reporting on everything in the data hub. These can help with understanding the housing landscape and affect decision making.
The reporting tool was still in the design phase when the fellowship ended, but the following designs give an idea of what the potential could be.
Note that all the data in the following screenshots is completely made up and is for demonstration purposes only.
Heat maps comparing the supply of affordable housing to the demand can help determine where investments should be made.
A trend line of overall supply of housing vs demand can help understand how it has changed over time.
A breakdown of how many units are supplied based on funding sources can help understand where affordable housing is coming from.
Conclusion
After completing the data hub and portal, we moved on to the final phase of our fellowship — building a housing search application. (A full account of that application can be viewed here)
At the end of the fellowship, the City of Austin took ownership of hosting and maintaining this data hub and portal. We still stay involved and up to date on how the project is going. The data hub and portal are currently being used by multiple agencies in Austin and the data will be used to help verify the city is following through on strategic initiatives and goals for affordable housing.
