Inspiration
The inspiration for this project came from understanding how the housing market behaved before, during, and after the financial crisis, and whether that behavior differed across regions. While aggregate trends often suggest a uniform recovery, we were interested in exploring whether all regions experienced the same trajectory or if structural differences existed beneath the surface. The goal was to move beyond simple trends and uncover hidden shifts in borrower behavior, market structure, and regional recovery patterns.
What We Learned
Through this project, we learned how to transform raw financial data into a coherent narrative using visual analytics. Specifically:
- How to use indexing to normalize time-series data and enable fair comparisons across regions
- How shifts in loan purpose mix and income distribution reflect deeper structural changes
- How to use scatter plots with reference lines to evaluate recovery at a granular level
- How secondary market dynamics and lender concentration influence outcomes
A key takeaway was that similar top-level trends can hide very different underlying realities when examined more closely.
How We Built the Project
The project was structured as a multi-act narrative, where each visualization builds on the previous one:
- ACT 1 (The Hook): Established baseline trends using indexed originations to compare regional growth
- ACT 2 (Behavioral Shifts):
- Analyzed changes in loan purpose mix
- Examined borrower income distribution
- Analyzed changes in loan purpose mix
- ACT 3 (Plot Twist): Compared 2007 vs 2017 MSA-level volumes using a scatter plot with a 45-degree reference line
- ACT 4 (Market Structure):
- Investigated shifts in the secondary market
- Analyzed lender concentration
- Investigated shifts in the secondary market
- ACT 5 (The Verdict): Classified markets into quadrants based on growth and recovery patterns
Tableau was used to build all visualizations, apply calculated fields, and organize the story into a structured flow.
Challenges Faced
One of the main challenges was maintaining consistency across multiple visualizations while still highlighting different aspects of the data. Early versions felt fragmented due to inconsistent colors, chart types, and layouts.
Other challenges included:
- Managing sorting issues for categorical fields like income buckets
- Handling legend complexity when combining multiple measures and dimensions
- Avoiding clutter while still conveying meaningful insights
Balancing simplicity with depth required several iterations. The final result focuses on clarity while preserving the core analytical insights.
Log in or sign up for Devpost to join the conversation.