Credit Where It’s Due

Inspiration

We were inspired by the scale of the problem hiding behind "simple" credit report mix-ups. A single data error like someone's report getting tangled up with a stranger's information can quietly block someone from renting an apartment, financing a car, or getting a job. When we dug into the Consumer Financial Protection Bureau (CFPB) complaint data, we found over 18 million complaints have been filed in the last 15 years, and the volume is exploding: complaints have grown roughly 5000% since 2011, with 12.5+ million of those specifically about credit reporting. Meanwhile, the CFPB itself has faced employee cuts and a backlog of 16,000+ cases, with a reported 22% backlog rate as of mid-2026. We wanted to build something that could help this broken system catch up.## What it does

Visualizes complaint volume over time, showing the sharp recent spike in credit-reporting complaints.
Maps complaints per 100K people by state, revealing that complaints are disproportionately concentrated in the Deep South (Georgia, Alabama, Florida, Louisiana, Mississippi, and neighboring states).
Breaks down how major credit bureaus respond to complaints
Surfaces the most common issues and sub-issues, with "information belongs to someone else" and "incorrect information on your report" making up over 70% of complaints.
Proposes three concrete solutions: a centralized data-transference system to help small bureaus exchange records in a unified format, a categorization system to rate complaints by urgency and cause, and an AI-driven anomaly detection system to flag companies with disproportionate complaint concentrations. ## How we built it We pulled CFPB complaint data, and normalized location with U.S. Census Bureau data. We parsed through the data, and used tableau for project visualizations, and solution ideas. We then researched how credit bureaus currently implement various systems, found where we could add support, and checked for implentation methodologies. ## Challenges we ran into
Small CRAs don't have the same standardized reporting infrastructure as the Big 3, making a "one-size-fits-all" data format hard to design.
Untangling which patterns in the data reflected real company issues versus normal regional or seasonal variation took several rounds of white-boarding, research, and reframing. ## Accomplishments that we're proud of
Turned a sprawling, multi-million-row government dataset into a clear, interactive story about a real, human problem.
Identified a concrete and unexpected geographic pattern (the concentration of complaints in the Deep South) that could inform where resources are needed most.
Designed a solution that balances AI efficiency with human oversight, rather than proposing full automation. ## What we learned
How the CFPB complaint pipeline actually works, from initial complaint to company response to resolution.
The technical realities of credit bureau data standards, like the Metro 2 format, and why standardization across agencies is so hard.
How anomaly detection can be used responsibly — flagging real problems without penalizing companies for normal seasonal or market fluctuations. ## What's next for Credit Where It’s Due Next, we want to build out the centralized data-transference and test it against real CRA data, as well as refine our urgency categorization program. Long-term, we'd like to partner with smaller CRAs and consumer advocacy groups to pilot the system.