wandr

Inspiration

Social media started as a way to connect. Now Instagram and TikTok optimize for likes, followers, and reach, so people perform instead of share, and ads fill the screen. Location features don't fix this: Snap Map shows where your friends are right now, not what happened there. Our memories end up buried in camera rolls, cut off from the places that gave them meaning.

We were inspired by "Crying at Columbia," an anonymous map of where students cried on campus. We also took inspiration from our new-member challenge Slack channels, where scrolling through everyone's submissions felt more connected than any feed. On a campus like Berkeley, thousands of students have their first day, worst night, and best friendship in the same places every year, and none of it is saved anywhere.

What it does

Wandr is a map of UC Berkeley made of memories. Do. Pin. Wandr.

1. Do: Live your life, or pick a Memory Maker challenge with friends.
2. Pin: Leave a memory at the exact spot it happened, with a photo (required), a caption, and a vibe. Everything is anonymous.
3. Wandr: Explore the map and discover other students' stories.

No likes, no feed, just Berkeley.

• Leave a memory anywhere. Drop a pin anywhere on campus, Northside, Southside, or Downtown, and the place name fills in automatically.
• Explore the map. Drag around freely. Pins cluster into bubbles that grow where lots of people were, and you can tap a pile to see everyone's memories from that spot.
• 🎲 Surprise me: fly to a random memory.
• 📍 Near me: see what other people remember near where you're standing.
• 🔒 Time capsules: memories you can only open when you're physically at the location.
• Memory Maker: to keep memories, you have to make them. There are three kinds of challenges:
  • Always on: swim in the Berkeley Law fountain, study 3 hours in Moffitt, study at the top of the Campanile, get Top Dog, spot a squirrel and name it
  • Limited time: get a picture with Oski on Cal Day, with a live countdown
  • Community: students post their own challenges anonymously
  • Finishing a challenge lets you pin the memory with a 🏆 badge, and a challenge filter shows every memory made from it.
• Anonymous, no virality. No usernames, profiles, likes, follower counts, or leaderboards, so there's nothing to perform for. You're there for curiosity, sharing, and exploring.

Demo use case: During midterm season, dozens of students pin memories at Moffitt, like a laptop at 4%, 2am snacks, or crying in the bathroom and then going back to finish. Moffitt becomes one big glowing pile of pins. Tap it and you see everyone's version of the same week, and you realize you're not the only one.

How we built it

• A web app in plain HTML, CSS, and JavaScript, built with AI-assisted coding (Claude)
• Leaflet and Leaflet.markercluster for the interactive map and the clustered "memory piles"
• OpenStreetMap data and Nominatim reverse geocoding, so any spot you pin gets a place name
• Browser Geolocation API with haversine distance math for "Near me" and the time capsule unlocks, plus a fallback that puts you next to Moffitt if GPS is denied
• Client-side image compression on a canvas, plus 128px thumbnails so the map stays fast
• A content pipeline: memories and challenges live in a spreadsheet, and a Python script turns it into the app's data
• About 170 preloaded pins built from our team's real photos, including a real study timelapse at the Campanile
• An animated opening screen with an original Campanile illustration and Berkeley icons that scatter from your cursor and pop into pins
• Figma for the future vision

Challenges we ran into

• Scope. We had far more ideas than time: Wandr Trails, navigation ("walk here"), privacy and close-friends controls, filters, and a leaderboard. We cut the leaderboard because it contradicted our no-virality idea, and we cut navigation because students already know campus.
• Figma vs. working code. A mentor told us a Figma-only submission was allowed but would score low on technical complexity. We chose to ship a working app and keep the creativity and polish.
• Maps and APIs. Getting a map running, plus clustering and reverse geocoding for pins placed anywhere, took real debugging.
• The image gallery. We had to fill the map with real photos matched to captions while keeping it fast, which is why we built the thumbnails and the spreadsheet pipeline.

Accomplishments that we're proud of

A working, demo-ready app with a full map from day one. It feels like Berkeley, and it stays true to our idea: it's built for connection, not performance.

What we learned

Deciding what not to build matters as much as what you build. We also learned that removing metrics like likes and followers changes what people share, and that making a demo relatable (midterms at Moffitt) beats adding more features.

What's next for Wandr

• A shared database with Berkeley email login, with posts still shown anonymously
• Reactions with emojis, and later custom stickers
• Close-friends vs. public sharing
• Reporting and moderation for community challenges
• "On this day": memories left at a spot on this date in past years
• Expanding to other campuses and cities

Built With

• css3
• figma
• geolocation-api
• html5
• javascript
• leaflet.js
• leaflet.markercluster
• nominatim
• openstreetmap
• python