Leading Light – Startup UX from Concept to Prototype

Challenge

Navigation apps optimise for time. They will route you down the fastest street at 11pm without any notion that the street is a factor in how the walk feels. For a lot of people — women, students, night commuters — that gap gets filled informally: texting a friend on arrival, taking the longer lit route, sharing a location before setting off.

Those habits are a product requirement that nobody had built. The question was whether safety could be the routing criterion rather than a badge bolted onto a map, with no funding, no proprietary data, and a team that had to build what it designed.

Solution

Lean UX, starting with more than ten interviews with urban pedestrians about how they actually decide a route — not what they would rank in a survey, but what they did last Thursday night.

Journey mapping turned that into a three-phase model, and the model became the product's structure.

Designs iterated from paper prototypes through wireframes to a mobile-first Figma interface, each round tested and revised. As the sole front-end developer, I built the working prototype in HTML, CSS, JavaScript, and React.

Decisions Worth Defending

• Reporting was designed for a stressed thumb. Icon-based, minimal taps, no free-text requirement. A reporting flow that assumes a calm user gets no reports from the situations that matter most.
• Danger zones glow rather than outline. A hard boundary makes a claim about a street that community-reported data cannot support, and it stigmatises specific blocks. A soft gradient communicates uncertainty honestly — the visual treatment is an epistemic decision, not a style one.
• Emergency contacts are chosen in advance. The moment you need the feature is the moment you cannot scroll a contact list. Configuration belongs in the calm phase.
• The three-phase model was used as a filter. Features that did not belong to prepare, navigate, or recover did not go in. On a team with no funding, the scoping discipline was the resource.

Outcome

The prototype tested well and drew interest from campus safety advocates and local safety groups. It never shipped, for three reasons worth stating plainly:

• The data was not there. Reliable safety data proved difficult to source and verify — and a routing product built on unverifiable inputs makes promises it cannot keep.
• Investors saw it as replicable. The judgment was that a large mapping platform could add the feature, which is a defensible read of the market even if it was not the whole picture.
• It needed critical mass it could not fund. Community reporting has a cold-start problem: the product is thin until enough people use it, and people do not use a thin product.

I keep this one in the portfolio precisely because it did not ship. It is the clearest example I have of full-cycle work — research, model, design, build — and of the more useful lesson, which is that a validated user need is not the same thing as a viable product. The interviews were right. The emotional model was right. Neither of those things solves a cold start or conjures data that nobody collects.