Download and Share the Report PDF!

Our work

Since the City of Detroit launched the Open Data Portal and GO DATA policy initiative in 2015, the Detroit Digital Justice Coalition (DDJC) and Detroit Community Technology Project (DCTP) have been exploring how to advance equitable practices for collecting, disseminating and using open data. By equitable practices, we mean accountable, ethical uses of public information for social good that actively resist the criminalization and surveillance of low income communities, people of color and other targeted communities.

To understand what these equitable open data practices could look like, we asked a broad spectrum of Detroit residents to consider the potential benefits and harms of various data sets currently available on the City's Open Data Portal. From there, we assessed what actions can be taken by the City to maximize benefits and minimize harms, and investigated open data policies and tools in place in other cities that model our vision for data justice.

The guidelines we offer below are the outcome of our research and outline how the City of Detroit’s Department of Innovation and Technology (DoIT) and other stewards of public information, in Detroit and beyond, can adopt and implement equitable open data practices.

We expect our recommendations to prompt critical conversations and questions, guide feature enhancements to Detroit’s data portal and inform policy provisions to improve how institutions collect, disseminate and use open data. We’re also eager to share specific takeaways from our research and resources for adapting our strategies.

Guidelines for Equitable Open Data in Detroit

Protect the people represented by the numbers

Issue:

While open data positively boosts government transparency, it can also be a source of fear and harm to residents. While facilitating data justice scenarios, we heard deep-rooted concerns about how the Improve Detroit Open Issues queue can incite property swindles, like a metal scrapper pinpointing an abandoned vehicle, how Blight Tickets can reinforce “bad” and “undesirable” neighborhood reputations and influence property values or red-lining practices, or how a randomized block-level location of a Crime Incident isn’t random enough in Detroit’s most desolate areas to truly protect from re-victimization, stigmatization or further harassment.


Guideline:

As the City continues to publish data, we ask for security audits to be performed and their outcomes shared by a third party, including penetration testing to expose potential vulnerabilities in existing data processing methods for anonymizing or randomizing personal data for release.

Don’t retain personal information tied to accessing City services

Issue:

We know that our personal information is often collected and stored when we apply for or access City services, such as entering a license plate number at a Park Detroit meter. While many privacy policies specifically promise to not share personal data with third parties, we want the City to go a step further and implement a provision to not retain any personal information associated with an application for City services. This would prevent the accidental loss or unauthorized use of personal data in ways that could harm or target residents.


Guideline:

New York City exemplifies how to roll out a successful Municipal ID program while pledging to not store cardholder’s personal background information, a policy that both safeguards the personal information of their cardholders, who are more likely to be low income, experiencing homelessness, undocumented, or people of color, and illustrates their commitment to Sanctuary City status.

Publish data about all City services, even for privatized “public” services

Issue:

The mission of the portal is, “to increase public access to valuable data and information concerning City government operations and service delivery.” But, what about crucial services in Detroit that are now administered by government contractors, private corporations and quasi-public entities, like the Public Lighting Authority, Great Lakes Water Authority or Detroit Economic Growth Corporation? These entities are responsible for data sets that we found residents often cite as most important to have access to, such as utility shut-offs and development projects, yet are not subject to FoIA laws.


Guideline:

The portal should provide basic contract information about all service providers at minimum, and ideally additional data about their specific operating activities. For example, what vendor(s) does the City contract to remove garbage? Where is that vendor located, what is the timeline and total amount of their contract? The Sunlight Foundation offers legal language for a provision like this to ensure “government decisions to employ outside contractors do not result in the public losing access to its own information.”

Prioritize the release of new datasets based on community interest

Issue:

Currently, two obvious ways exist for residents to request public records not yet published to the data portal: through the portal’s nomination form and through Freedom of Information Act (FoIA) requests. The first option requires internet access and appears to be a slow process, as only a single nomination from over 30 suggestions has been approved since launching. The second option, FoIAs, are also notoriously cumbersome and can be costly. Yet, these two mediums remain crucial ways for the City to gauge community interest and data needs.


Guideline:

FoIA requests and their responses, as well as approved nominated data sets, should be posted to the portal, to both boost relevant data sets for users and save internal resources in answering duplicate requests. We outlined our favorite ways that other cities are accomplishing this in our policy research.

Increase transparency around how data sets are defined and processed

Issue:

While we appreciate open data efforts, many data sets are not so easy to decipher. Participants in our data justice focus group often pointed to an example of confusing language in the Blight Violations dataset. How can we begin to use or analyze these data if we don’t know how the City defines “blight”?


Guideline:

All data sets published on the Portal should include field definitions, collection methods, especially for anonymizing or randomizing personal data, and basic definitions of policy-specific or technical language. The DPD: Citizen Complaints dataset exemplifies useful, thorough documentation by offering a description, contact information and field definitions, yet we’ve found many other data sets in Detroit’s portal to be lacking these metadata. Beyond Detroit’s portal, we also looked to national cases where open data are being used for predictive analytics, like predictive sentencing in Philadelphia as expanded in our policy research. The algorithms and analysis methods used are often inaccessible to the public and not subject to community oversight, yet can be encoded with implicit biases that affect how different factors are weighed. We ask that transparent practices apply to all open data definitions, methods, analysis and algorithms.

Engage residents offline about open data

Issue:

In order to build awareness and open data literacy in Detroit, it’s important to consider digital inequality and limited access to broadband internet. Through our data justice surveys, we found that most residents first learn about open data and the portal at events or workshops, but these can be infrequent and hard to access.


Guideline:

In addition to the online tutorials DoIT currently offers for learning to use the portal, we need an engagement plan that emphasizes community dialogue and in-person training sessions. We’re sharing strategies we’ve found successful through organizing Data DiscoTechs, like partnering with local recreation centers and developing popular education methods to interactively learn about open data, as a starting point.

Share what’s coming next

Issue:

While it’s simple to see how the Portal has grown since its launch through the Public Dataset Inventory, we also want to know what’s coming next so that we can plan community research and organizing initiatives.


Guideline:

Once or twice a year, DoIT should publish their key goals and activities, sharing with the public which dataset releases, portal feature improvements and public events to anticipate. Similar to the NYC Open Data Portal, the header of Detroit’s portal should rotate to highlight features, identify the purpose, advertise new releases and offer quick links to things that matter most to Detroiters.

Connect with us!

We anticipate that our recommendations, and the research that has shaped them, will continue to grow and evolve as we continue organizing around equitable open data.

We welcome your feedback, ideas and links to additional resources for growing this work. Reach out to us by email at communitytechnology@alliedmedia.org or submit an issue to our Github repository.