Lessons Learned from FEAT State-Scaling Efforts

From staffing and budgeting to disseminating insights based on housing loss analysis, the following lessons learned are intended to shed light on the process of scaling the Foreclosure and Eviction Analysis Tool (FEAT) and provide tips for jurisdictions interested in replicating these efforts.

Data Efforts Must Be Adequately Supported with Financial and Human Resources

The collection and use of eviction and foreclosure (EF) data requires significant staff capacity in the form of time, financial resources, and labor. Absent a mandate, funding is the most powerful tool for cementing data use as a priority. Whether these efforts are undertaken to the degree required, and on a regular and systematized basis, often hinges on whether there is adequate funding and staffing provided for these efforts.

Creating Eviction and Foreclosure Data Dashboards Requires Multidisciplinary Teams

The quality and complexity of EF data requires both housing expertise and data proficiency. Undertaking meaningful analysis of this data will require a multidisciplinary team with a range of skills, including data management, geospatial analysis, information technology (IT), and data visualization, as well as deep knowledge of the legal processes around eviction and foreclosure. Assessing the skill sets on the team—whether those are in-house or through partnerships (e.g., with local legal aid providers)—and filling any gaps in expertise is essential.

If Data Pipelines Are Not in Place, Accessing the Data Can Cause Significant Delays

Even under the best circumstances, requesting and obtaining court data is a lengthy process. That is because in most jurisdictions—at least the ones New America has worked with—access to this data involves several steps (typically the submission of a bulk data request followed by entry into a data-sharing agreement with the court) that are not bound by delineated timelines. Processes that in theory could take a matter of weeks typically stretch to several months. A data collection effort must assume roadblocks and significant delays and have ample staff capacity to engage consistently and persistently to overcome them.

Gather the Insights of Target Dashboard Users, Early and Often

Gathering insights from the range of potential users is a critical first step in developing the dashboard. While there will be similarities in the types of information dashboard users are most interested in (e.g., the location of eviction hot spots), other needs vary according to local pressures, policies, and even geographic features.

Hosting community sessions and running a user-centered dashboard design process that involves the range of potential users will help prioritize data analysis that meets the needs of local communities and integrate design elements that ensure dashboards are presenting insights in the most useful way.

Be Aware That User Needs Will Not Align Perfectly with Data Access and Quality

User needs will not align perfectly with what is feasible, given the limitations around data access and quality. For example, some users may be interested in insights that require obtaining data beyond what can be accessed from courts, such as the prevalence of out of state landlords among top evictors or whether eviction and foreclosure hotspots are coinciding with flood zones and other climate vulnerable parts of their community. It is still important to start with blue sky suggestions as these can be prioritized for future consideration.

Understand That Courts Often Lack the Capacity and Incentives to Provide Public Court Data

The reason it is so difficult to obtain eviction data is that it’s time consuming and tedious for court staff to provide it, and there’s often no mandate, nor much of an incentive, for courts to share data with non-court users. Even when a court has a process for bulk data requests, court staff are typically provided little time and resources to fulfill these requests. In courts without rules on the books around provision of data it is in fact not a part of a court’s job. For these reasons, providing data in the particular format or on the particular timeline that the public needs to conduct meaningful analysis often falls to the bottom of the long list of a court staff’s tasks. Understanding this dynamic is critical to working within the system to align incentives and access the necessary data. Here are some strategies that have worked well.

• Find and partner with the local data guardian: In jurisdictions where there are not clear data-sharing processes, find the individual who is either formally appointed to manage data-sharing or who has informally taken this role upon themselves. Advocating for your data needs with the right individual is key in both formal and informal data-sharing arrangements with court systems.
• Be clear about the benefits of data transparency for courts: Courts nationwide are under pressure to cut caseloads and processing times. For that reason, it’s helpful to make the case that improved data access can help decrease evictions and foreclosures, and therefore decrease caseloads.

Align Incentives for State and Municipal Leaders to Unlock Court Data Access

City, county, and state officials can be powerful allies in unlocking access to court EF data. In addition to believing that better data will help decrease evictions and foreclosures, many municipal staff can make the case to their colleagues and superiors if EF data efforts are tied to other major initiatives or incentives.

• Appeal to innovation: Local government staff are motivated by the recognition of being innovators and leaders. For this reason, we have found it effective to tie data-sharing efforts to larger themes of “transparency” and “innovation” and to existing learning cohorts like the National League of Cities’ and Stanford Legal Design Lab’s Eviction Prevention Learning Lab.
• Appeal to inter-municipal competitiveness: Municipalities within a state are often in competition with one another, and a municipality is often motivated by being the “first” to adopt an innovation. For this reason, once a municipality within a state has agreed to share data, it is sometimes effective to share this precedent with other municipalities.

Municipalities Can Work Together to Open Data Doors

There’s no need to reinvent the wheel when it comes to data requests, structure, and format. Once a single municipality in a state has developed a process for requesting and sharing data, often through a data-sharing agreement, that process can be replicated in other municipalities. This also applies to the data fields and specifications within the data request; in fact, for purposes of standardization it is optimal for all municipalities to provide data in the same format.

Even “Clean” Data Will Require Significant Effort to Prepare for Analysis

Given the quality and complexity of eviction and foreclosure data, working with this data will require a multidisciplinary team with a range of data skills and housing expertise.

Raw data feeds from county clerks’ offices are often provided in pipe-delimited formats that contain different variable names than what data analysis tools like FEAT are set up to ingest. A team working with these files must have both IT skills and housing knowledge to correctly convert data from courts to a format that FEAT and other tools can read. It is helpful to have a counterpart at the court who is knowledgeable about and responsive to questions about how the data is generated, how it should be interpreted, and any limitations in its quality or comprehensiveness.

Pairing Data Analysis with Local Eviction and Foreclosure Expertise Is Essential

The complexity and variation in eviction and foreclosure legal processes across the United States, in combination with often poor-quality data, presents an opportunity for data inaccuracies and misinterpretation. Experts with deep knowledge of the relevant legal processes and housing policies and programs must participate in the EF data analysis process.

For example: Mortgage foreclosures involve multiple court filings over a many month period, and the process varies by state. It is essential to have a granular understanding of the local mortgage foreclosure process to pull out the correct court filing for inclusion into FEAT.

Simple Is Best

Remember that a dashboard containing multiple datasets and filters can become difficult to understand, and, therefore, be less useful. For most dashboard and data analysis, simplicity is best.

Developing a Dashboard Is 95 Percent of the Work, but the Utility Is in the Last 5 Percent: Converting Analysis into Insights

It is incredible how much work goes into being able to say that “Neighborhood x has the highest foreclosure rate in the city” or “Evictions are highest in month y.” Equally incredible: Many dashboard developers do all the work necessary to surface these insights but stop short of actually publishing them in plain text on the dashboard interface. Users must be able to easily find these insights without having to do a lot of additional investigation or digging. Therefore, regardless of how “user friendly” a dashboard is, it is essential that in addition to the data analysis and visualizations, a dashboard contains a set of “key takeaways” based on needs identified by users during the dashboard design process.

Invest in Communicating Insights: Just Because a Dashboard Is Released Does Not Mean It’ll Be Used

Dashboards must be continuously socialized, updated with new data, and adjusted in response to policy and legal changes and user feedback. In an ideal case, dashboard development should plan and budget for at least two more years of funding post-release.

Plan to spend significant time communicating about the existence and functionalities of a dashboard to local housing communities who can benefit from its use, even if users were consulted in the lead-up to dashboard development.

A suggested dashboard socialization process includes: a public release event; a mass email distribution; briefings and embargoed viewings for local housing reporters; individual briefings and demos for key users from government agencies, legal aid organizations, service providers, and advocacy groups; and written collateral on your website to accompany the dashboard itself.

Plan on (and budget for) repeating this process each time the dashboard is updated with new data.