Jessie Finocchiaro and the Conversations with Practitioners Working Group
Katie Bridges is a senior business intelligence analyst for the city of Boulder, Colorado, who has a passion for using data to inform decision making in the public sphere. Our Mechanism Design for Social Good (MD4SG) working group, Conversation with Practitioners, had the pleasure of interviewing Katie. In this blog post, we record the main insights from the interview for our audience of researchers.
Focus on data literacy: is less sometimes more?
In addition to her professional role, Katie conducts training sessions for community members on data literacy. Communicating insights from data effectively is a skill which requires practice and creativity. Katie also helps local employees to thoughtfully develop questions that can be answered with data. “Oftentimes, people [just] want data for the sake of data.”
How can public sector institutions achieve accountability through open data?
Data communication in the public sphere has improved tremendously in the past decade, but there is still a long way to go. Many cities depend on budget reports and city council meetings as their primary method of data communication, which may not convey the information residents need. One way to improve data communication is through anonymized open datasets. (For example, Katie has worked on the Open Data Hub in Boulder.) Citizens can use open data to take their questions into their own hands.
What are the various methods of data collection used?
Data collection methods vary tremendously. For example, within the realm of public transportation and infrastructure, bike usage data might be collected by having bike counters observe different streets in a city, while curbside management data on COVID can be collected by city vehicles with license plate readers. Considering the how of data collection enables us to ask if these methods are sufficient or how they might be improved for a given question.
What data is collected? How do people know you’re collecting data on them?
In the case of the city of Boulder, outreach teams communicate data usage to residents in accessible ways, such as by handing out pamphlets. A racial equity program manager, neighborhood liaison, and communication liaison work together to ensure all voices in the community are heard. For example, this team collected paper surveys at neighborhood gatherings to guide the values for selecting a new city manager. It is a tremendous challenge to make open data available to everyone, such as residents without computer or internet access. The city aims to find creative ways to communicate their data, even to residents who have limited computer access.
How can you balance data privacy, equity, and open data?
The question of how to balance data privacy and equity is a difficult one without a complete answer. Katie gave an example surrounding inquiries about nuisance activity, loosely defined as anything that infringes on quality of life, such as noise violations, sidewalks being blocked, lawns not being mowed, among others.
“[Nuisance activity is] disproportionately reported against younger and lower-income folks. It is the city’s responsibility to ensure quality of life, so the city has data on these police calls and to which addresses enforcement teams were sent. There was significant pushback on making this data public, even in an anonymized form, due to the chance of retaliation against the presumed residents of these homes. However, federal obligations meant the data had to be made public.”
This type of issue requires careful thought and decision making to protect our most vulnerable members of our community.
What matters more to the public sector — prediction or analysis?
Due to resource constraints, the public sector mostly focuses on analysis rather than prediction. Agencies need resources to respond to predictions. “For example, suppose we hand a prediction about folks who might need help with food or housing to our Housing and Human Services team. Quite frankly, they often don’t have the staff to go out and intervene before it’s an issue.”
Although, some projects have started to use prediction, trying to predict where interventions might be necessary in the future requires building out a data warehouse that merges different data sources; this makes prediction relatively rare in Boulder as it stands; If we ask about the analysis to prediction ratio in another year or two, that answer might be a bit different, though. For example, in the past year, some work on predicting COVID outbreak locations was done to inform the location of COVID testing and vaccine distribution sites.
Interested in getting involved?
“It’s really important to meet people where they are. To me, that can be really challenging, because you spend years of your life studying mathematical models, and then the person in front of you often isn’t going to understand that model, so having a translation mindset is crucial. When I present to transportation folks, they don’t need to know how I developed a model, but they do need to know the outcome and its translation to their day-to-day work.”
Katie suggests thinking about city employees and their day-to-day: think about processes they perform often that are prone to mistakes. “To me, it’s about thinking about the practitioner’s experience and thinking about how technology can augment their work without taking away from the practitioner’s expertise.”
We would again like to thank Katie for her insights, as well as the members of the Conversations with Practitioners for their engagement and thoughtful questions.
The interview with Katie was led by Jessie Finocchiaro, and this blog post was edited by Jessie, Jeremy Vollen, and Matthew Olckers.
