Targeting for Impact: Helping a Federal Agency Mine “Dark Data” to Improve Citizen Outcomes

A Federal regulatory agency turned to Censeo to develop a fact-based and nuanced understanding of issues facing different population segments, enabling it to target its outreach & education and enforcement activities for maximum impact and improve citizen outcomes.

Situation

A Federal regulatory agency was under pressure to improve the efficiency and effectiveness of its consumer outreach and enforcement activities and drive better citizen outcomes. Agency leadership felt that current outreach efforts were based on “legacy assumptions” not attuned to the latest issues confronting citizens, and that enforcement actions were often reactive, instead of being informed by trends in the issues facing consumers of the services the agency regulated. The agency engaged Censeo to develop a data-driven and fact-based approach for targeting its outreach & education and enforcement activities.

Approach

The Censeo team recognized that the core problem was lack of localized and real-time time insight into consumer issues. If the agency were to have visibility into the specific issues facing citizens in a given zip code or county for instance, and the trends in these issues over time, it could target its outreach efforts toward educating consumers around their rights and mechanisms to address these issues and avoid them in the future, and simultaneously launch proactive compliance and enforcement actions targeting the actors causing these issues. Simply put, they could proactively educate potential future victims and rapidly pursue enforcement actions.

Consequently, the team’s first priority was to develop a dataset that could provide this insight. The Agency’s call center already maintained a database of call records, including transcripts of citizen complaints. However, this was a classic example of “dark data” – data that could not be immediately mined for insight – since the key information was buried in unstructured text fields containing the call transcripts. The Censeo team deployed advanced text analytics algorithms on this call transcript data to categorize each call by the specific issue(s) covered during the call. Next, the team combined this categorized database of call records with external datasets (e.g., U.S. Census Bureau data) to generate a final analytics-ready dataset that could be mined for localized insight into citizen issues.

Finally, the team packaged these insights into intuitive and user-friendly interactive “issue heat map” dashboards (see graphic) that decision-makers used to inform education and outreach plans and enforcement actions at the region, state, and county levels in the near-term and policy actions in the longer-term, thereby driving improved consumer outcomes.