Data & decisions / 07
public mortgage data
Mortgage market intelligence
Every U.S. metro scored each month from Census, HMDA and FHFA data, ranked by the strategy weights you choose.
The decisionWhich U.S. markets a mortgage business should grow in, and why.
01 / Overview
The problem, and the approach.
AThe challenge
Decisions about where to grow often rest on anecdote or a single metric, while the public data that could inform them sits in three agencies’ formats.
BThe approach
A scheduled pipeline pulls Census, HMDA and FHFA data each month and scores 380+ metro markets on five pillars. You set the strategy weights and the rankings rebuild; PCA and K-means group markets into named archetypes, lender shares show who already competes, and an anomaly score flags markets that stand out.
02 / The decision journey
5 decisions, one story.
Each stage asks one business question and answers it with a model you can open and change. Any card opens that stage of the live demo.
-
01
Score
Where is the mortgage market opening up?
Five pillars, your own strategy weights, rankings that rebuild on the spot.
Open this stage
-
02
Archetypes
Which markets behave alike?
PCA and K-means group metros into archetypes with plain-English names.
Open this stage
-
03
Lenders
Who already competes there?
HMDA lender shares and competitive openness, market by market.
Open this stage
-
04
Brief
What should a leader know about one market?
A grounded brief written from that market’s own numbers.
Open this stage
-
05
Signals
Which markets look unusual?
Gaps between demand and lending, and an anomaly score for markets that stand out.
Open this stage
03 / Under the hood
How it’s built.
Algorithms
- PCA
- K-means
- Composite scoring
- Anomaly detection
Stack
- Python
- pandas
- scikit-learn (PCA, K-means)
- Scheduled pipeline
- Plotly
- Streamlit
Methods, in order
- Scheduled pipeline
- Composite scoring
- Clustering
- Lender benchmarking
- Anomaly detection
- Grounded Q&A
Data
Public data only: Census ACS, HMDA and the FHFA house price index.
04 / Outcomes
What it changed.
- 01
380+ U.S. metro markets scored every month without manual work.
- 02
Market archetypes with plain-English names a leader can use.
- 03
A grounded brief for any market, written from that market’s own numbers.
Results come from synthetic or public data with known answers. They show the method, not a client result.
05 / Try it
The live demo.
The real app, running on synthetic data. Change the inputs and watch the decision move.
Interactive, synthetic data. If nobody has used it for a while it takes about 30 seconds to wake.
Open in a new tab ↗06 / Related
Keep exploring.
Data & decisions / 08
Mortgage product strategy
Where should a lender expand? Public market data, weights you set, a stress-tested ranking and an executive brief that writes itself.
View project
Interactive lab / 01
Lending decision lab
A fictional credit union and 50,000 synthetic members, followed through nine connected lending decisions: who to approve, what to charge, who will leave, and whether it is all fair.
View project
Applied AI / 04
Credit policy assistant
Answers from a lending policy manual with every claim cited, and weak matches declined instead of guessed.
View project
Open conversations