A neighborhood research tool I built for my own relocations, as my wife and I moved to a new state every three months for her optometry rotations.
Personal-use prototype · Not publicly hosted. Watch the recorded walkthrough.
StatusPersonal-use prototype
Primary userMe
Built withClaude Code
My role
Product idea, experience design, development with Claude Code, and testing through my own relocation research. Built with Next.js, Google Maps Platform (Places, Geocoding, Routes), and city police open data.
01
The problem
My wife’s optometry rotations meant moving to a new state every three months, sometimes splitting our time between two cities. Each move meant finding an area that fit her commutes and our everyday needs.
I kept switching between maps, Claude, local research, and furnished rental sites. I built blocky to bring that search together.
02
What I built
Enter your work addresses, rank what matters to you, and compare neighborhoods on one map.
Compare commutes to up to three work addresses.
Reorder your priorities—commute, safety, and nearby amenities—and see neighborhood grades update.
Check a rental address for estimated rush-hour drives and places within a mile.
View a crime heatmap with clear labels for public data and estimates.
I used earlier versions during moves across Kansas, Missouri, Florida, California, and Pennsylvania.
03
Key decisions
Build around what decided our moves. I started with mosques and halal food, then expanded to commutes, safety, gyms, and coffee as I used it.
Show where the data ends. I used publicly available city crime data. Where it wasn’t available, I showed estimates in a faded style and labeled them.
Find the neighborhood first. Narrow the search before comparing individual rentals.
Guide the user one step at a time. Initially, the left panel and map icons appeared before you entered an address. I changed the flow: enter a work address, see recommendations on the left, then smoothly load the map. This kept the screen from feeling overwhelming.
Add personality. Blocky greets you, winks when you enter an address, and hops into the corner as the map loads.
04
How it was built
I designed the product, built it with Claude Code, and tested it against my own housing search.
Check the recommendations
Missing crime data gave one area a misleading safety score. A commute-scoring issue also favored suburbs far from two clinics. I caught and fixed both by checking the output against real places.
Speed up the map
Filters lagged while the map drew about 1,600 pins and 1,600 colored squares. I reduced the pins to about 230 and combined the heatmap into one image. Filters then felt near-instant.
Reuse the data
Caching the area’s data let me update rankings without repeating the same lookups, reducing waits and API costs.
05
Results & what’s next
Results
A working prototype I used for my own moves. The Philadelphia-area dataset includes about 660 gyms, 360 coffee shops, 85 mosques, and 52,000 police reports from a year of data.
What I learned
Finding the right area was only half the job. I still had to search Furnished Finder separately to see what was available. Connecting neighborhoods to available rentals would make the tool more useful.
I also learned to make gaps in the crime data clear instead of presenting every area with the same confidence.
What I’d test next
Show available furnished rentals on the map.
Add more categories for traveling healthcare workers.
Request crime data from suburban police departments to fill coverage gaps.
