Full-stack web app · 2026
Voteometer
A decision-support tool that blends how much you like each candidate with how likely they are to win, producing a single "Power Number" for who to vote for.
Overview
Voteometer is a political decision-support web app that helps voters choose the best candidate by combining personal preference with real-world electability into a single score — the Power Number. Instead of asking “Who do I like best?” or “Who can actually win?”, Voteometer answers both questions at once and surfaces the candidate with the best shot at delivering your preferred outcome.
Problem
Voters in multi-candidate primaries often face a dilemma: their favorite candidate may have little chance of winning the general election, while a more electable candidate may not align well with their values. Existing tools address preference or electability in isolation but not both together.
What I Built
- A rating system where users score each candidate on a -10 to 10 preference scale
- Head-to-head matchup questions to surface relative preferences
- Integration with Polymarket to pull live market probabilities for general-election matchups
- Multi-candidate primary modeling with automatic question generation
- A Power Number formula that weights preference against win probability
- Interactive Chart.js visualizations showing score breakdowns and comparisons
- Full deployment on Vercel
Challenges
The trickiest part was designing the Power Number formula so that it felt intuitive — a candidate you rate 10/10 with a 5% win chance should score lower than one you rate 7/10 with a 60% win chance. Tuning the weighting to feel fair across a range of scenarios required several iterations of testing and user feedback.
Outcome
Voteometer is fully live and usable. Building it deepened my understanding of TypeScript, data modeling, and how to translate an abstract idea into a working product with a clean, usable interface.