Whspr: Crowdsourced urban intelligence platform for women

Women already share knowledge about places all the time. Before going somewhere new, they'll text a friend, check a neighborhood subreddit, or read reviews to get a sense of what a place is actually like. But that knowledge is scattered and it disappears fast. It lives in group chats and comment threads, and none of it is anywhere you can actually find when you need it.

There's no interface designed to support collective sensemaking around women's perceived safety in public spaces.

I began by researching why women still resort to the whisper network in the age of apps, then went out and asked them. Three patterns kept coming back.

I set out to build a safety tool. The research told me I was solving the wrong problem.

Women weren't asking for another app to warn them about danger. They already had that, and it wasn't working. What they wanted was the thing they were already giving each other in group chats: a sense of what a place actually feels like, at a certain time, for a certain kind of person.

So I stopped designing a safety app. Women already notice things, tell each other, and warn each other. What they don't have is somewhere for all of that to live after the group chat moves on.

Four decisions, each one traced back to the research.

When I asked one woman whether other people's experiences change where she goes, she said, “It depends what the experience was, if it was a staff member that made the experience bad, or if it was just someone who was there.” That difference is exactly what a star rating throws away, and it's what the prompts are there to capture. The cost is that posting takes longer, so whether people keep doing it at scale is the first thing I'd test next.

The design system

Whspr runs on a dark, calm interface, built to feel like a quiet resource rather than an alarm. A Midnight base, a single Signal Amber accent, and DM Serif Display paired with DM Sans.

This is the working prototype, so go ahead and use it. Search a place, read what other women said about it, or post something yourself.

“The interface had to do two things at once: let someone post what a place was like, and let someone else read it to decide whether to go. The whole design keeps those two jobs working together without letting either one flatten the other.”

You search a place and see what other women experienced there, tagged with when they went and what kind of visit it was. There's no score and no crime feed, only people who were actually there.

To test the concept I took it out around Washington Square Park and the West Village. I talked to 10 women across 7 short conversations, plus 2 long-form interviews. Each street conversation followed the same order, suggested by Professor Bloom: validate the problem first, show the prototype second, ask about use last. That way people named the problem in their own words before they ever saw my solution.

The premise held up. Women already share this stuff, so I didn't need to invent a new behavior, just give it somewhere to go. And the problem was never a lack of information. People just didn't trust where it came from.

The two who said no were right about something. One asked straight out why she'd use this over reviews. The other doesn't own a smartphone, but while ruling herself out she described exactly who Whspr is for: someone new to the city, without a list of places she already trusts. Between them they scoped the product better than I had. Whspr is for people going somewhere they don't know yet.

“If you've never been in an environment like that, how do you trust your instinct?”

Interview participant

From places to routes. Nobody was asked what else Whspr should do. Getting home came up anyway, from more people than any feature I actually designed. For one woman, the trip home mattered more than anything about the venue. A Whspr layer over the map, where women could leave and read feedback along a route, is the most-requested thing I didn't build.

Verification and inclusion. The verification I designed confirms identity but raises real questions I didn't resolve. Selfie and pose-based checks shut out low-vision users, and gender verification gets complicated fast for trans women, who Whspr is meant to include. I'd want to talk to both groups before building any of it.

Trust at scale. A platform built on contributions is only as trustworthy as the people contributing. Guarding against coordinated manipulation and astroturfing is something I'd need to design for before this could be real.

Women already know this stuff. The apps around them just don't treat it as real. Whspr made me notice how much the default choices, like a star rating or a crime map, decide whose experience counts. Saying no to those defaults was the part of this project I cared about most.