FairTenant

Tenants reliability stats
Comparison between legacy and new model

Inspiration

Tenant screening tools are meant to protect landlords from financial risk, but many rely on signals like credit history, employment length, or past evictions. These signals are easy to use, but they often reflect social and historical bias rather than whether someone can actually pay rent.

Landlords may not intentionally care about bias, they care about missed payments and vacancies. We were inspired by the idea that bias persists not because it’s fair, but because it’s assumed to be profitable. Our question became: what if reducing bias actually makes landlords more money?

What It Does

Our project compares two ways of evaluating the same tenant:

A legacy screening model that relies on traditional proxy signals commonly used today
A cash-flow–based model that focuses on affordability, savings, and income stability

For the same tenant, we show:

How each system scores them
The expected number of missed rent payments
The estimated financial loss

By putting these results side-by-side, we show how proxy-based screening can quietly reject or undervalue reliable tenants, while a cash-flow approach often leads to better financial outcomes.

How We Built It

We built a working backend that runs both models in real time.
The legacy model represents how traditional systems behave, while our new model intentionally ignores social proxies and looks only at financial reality.

To make the comparison fair, we clearly document assumptions and translate both scores into dollars lost, making the impact easy to understand.

Challenges We Faced

A major challenge was that legacy systems depend on information we intentionally don’t collect. Instead of hiding this, we modeled how those systems infer risk when data is missing. This helped us highlight how silence in data often gets filled with biased assumptions.

What We’re Proud Of

Showing that reducing bias doesn’t require moral arguments, it can improve profits
Turning abstract fairness concerns into clear financial consequences
Building a live demo that clearly shows how different assumptions change outcomes

Noise and Silence Connection

Bias often hides in what systems choose to listen to and what they ignore.
Proxy variables are noisy signals that obscure what actually matters, while genuine financial stability is often silent or overlooked.

Our project examines how eliminating noise and focusing on the right signals leads to more informed decisions.