Ripple

Inspiration

We wanted a way to test “fair” policies on real-looking populations without needing a whole data science team. Cities, campuses, and nonprofits often have spreadsheets of people, income, rent, commute, or energy use — but no quick way to ask: if we raise wages, or add a housing subsidy, or encourage EVs… who actually benefits? We built Ripple (our fairness-aware policy simulator) so anyone can upload a dataset and see how a policy change ripples through different groups.

What it does

Ripple is an interactive policy sandbox. You upload an Excel sheet of people/households, then:

Pick policy levers (minimum wage increase, housing subsidy, transit/EV incentive, education or green-jobs boost).
Ripple simulates how that affects income, rent paid, employment, commute, emissions, and even energy use.
It breaks results down by race/ethnicity, county/region, income bracket, and sector so you can spot uneven impacts.
It also tracks things that are usually hidden, like rent burden and average rent actually paid, so we can tell if a policy is good on paper but bad for low-income groups.

So it’s basically: Upload → Adjust policies → Get fairness and sustainability insights.

How we built it

React frontend for the main dashboard and cards (baseline metrics, post-policy metrics, group breakdowns).
Excel → JSON pipeline in the browser so a judge or policymaker can just drop in a sheet and we detect common column names (income, rent, race, county, energy, commute).
Simulation engine that, for every person/row, applies the chosen policies: updates income, recalculates rent and subsidy, changes commute mode if there’s a transit/EV incentive, and updates energy use.
Aggregation layer that recomputes group-level metrics (by race, county, income bracket) so we can say “this group improved X% from its own baseline,” not just from the overall average.
Monte Carlo-style runs (lightweight) to smooth randomness and give more realistic ranges.
Styling and visualization hooks so we can plug in better charts (Lorenz, before/after, equity plots) without rewriting the logic.

Challenges we ran into

Messy input data: different spreadsheets name the same column differently. At first that made whole groups show up as “Unknown,” which invalidated the equity view. We fixed it by checking multiple column-name variations everywhere.
Housing realism: if you only track “rent,” a wage increase can make the results look worse. We added rent actually paid (after subsidy) and rent burden to show the true effect.
Linking climate and equity: EV/transit incentives lower transport emissions but can change energy use. We had to track energy too so we don’t claim a win in one dimension and hide a loss in another.
Performance in the browser: running per-person simulations with several levers can get heavy, so we tuned the number of runs and how much we log.

Accomplishments that we're proud of

Turned a plain Excel upload into a full policy simulator in one flow.
Made the results equity-aware — you can actually tell which race/region/income group benefited.
Added practical metrics (rent paid, rent burden, energy use) that make the demo believable to real stakeholders.
Documented future visualizations (Lorenz curve, Sankey for mode changes, radar for policy comparison) so the project looks extensible beyond the hackathon.

What we learned

Data robustness matters more than fancy charts — if your app breaks on one column name, nobody trusts it.
Equity analysis only makes sense if you store per-group baselines first.
Policy levers interact (wage ↔ rent, EV ↔ energy), so it’s better to show multiple outcomes on one screen.
A clean, card-based React UI is enough to tell the story to judges quickly.

What’s next for Ripple

Add the visualization set we outlined (Lorenz, before/after income, equity dot-plot).
Let users save and compare scenarios (Policy A vs Policy B vs Policy C) on a single radar chart.
Add budget/cost estimates for each lever so you can say “best outcome under $X.”
Add light geo (by county/region) to show where the policy hits hardest.
Package the simulator as an API/notebook so other civic/hackathon projects can drop it into their stacks.