Consumer Complaints EDA (2011–2024)

Business Problem

Rising CFPB complaint volumes create reputational, compliance, and revenue risks for banks and fintechs. Without a clear view of spikes, firms struggle to prioritise fixes, allocate resources, and stay ahead of regulators.

Objective	Business Value
Early-warning radar	Detect product failures before they trigger fines or class-action suits.
Peer benchmarking	Compare complaint rates by product & state to expose outliers.
Root-cause targeting	Trace surges to specific workflows (e.g., credit-report fixes) and redesign them.
Resource allocation	Focus CX budgets on hotspots (FL, TX, CA) and high-risk categories.
Regulatory readiness	Give risk committees KPIs that pre-empt CFPB investigations.

Volume of CFPB complaints 2011-2024 — Figure 1. Volume of complaints over time.

Data & Methodology

We analysed 6.9 million CFPB complaints from 2011-01-01 to 2024-03-31, downloaded via the public API and processed in R. Below is the five-step workflow that ensures data quality and insightful visuals.

Ingest & type-casting — Load raw CSV and convert dates / factors with readr::read_csv.
Data cleaning — Drop duplicates and empty narratives (0.6%) using dplyr.
Date enrichment — Add Year and Month fields via lubridate for time-series aggregation.
Geo-join — Map two-letter states to FIPS codes (usmap, sf) for choropleths.
Visual EDA — Create heat-maps, bar charts and line plots with ggplot2 & leaflet.

Limitation: dataset only captures consumers who filed formal complaints; silent dissatisfaction is not recorded.

What I found

Key Insights

01 32% of all CFPB complaints filed since 2011 are credit-report errors — the single largest complaint category.

02 Complaints doubled post-2020 vs. the pre-COVID baseline, driven by pandemic-era financial hardship.

03 FL, TX & CA generate 27% of total complaint volume, reflecting both population size and financial-sector concentration.

Next Steps

Publish monthly credit-report KPI dashboard for risk teams.
Pilot chatbot resolution in hotspot states.
Apply NLP topic modelling to flag emerging issues.

Conclusions

The CFPB dataset highlights persistent issues in credit reporting and debt collection. Geographic analysis shows disproportionate complaint volumes in populous, financially complex states. The pronounced uptick post-2020 underscores how economic shocks translate into consumer-finance friction.

Credits

Analysis by Juan Camilo Sierra Escobar, December 2024.
Data sourced from the Consumer Financial Protection Bureau (CFPB).

Repository

View full code on GitHub

Explore the project

View on GitHub Back to Portfolio