DEV Community: Auu Cooper

I Built NabeWise Because “Best Neighborhood” Is a Very Suspicious Phrase

Auu Cooper — Wed, 07 Oct 2026 08:28:02 +0000

There’s a particular kind of internet research that starts with one innocent question:

“Where should I live?”

Twenty minutes later, you have apartment listings, a map, three neighborhood guides, and a Reddit argument where everyone involved seems personally offended by a bridge.

One person says an area is perfect.

Another says it was perfect ten years ago.

Someone else recommends a completely different city.

Helpful. We have successfully expanded the problem.

I’m building NabeWise to make the neighborhood part of that decision a little easier. It helps people explore areas in U.S. cities and compare things like rent, walkability, schools, safety, and commute.

The idea sounds fairly simple.

Then you try to define “neighborhood.”

We’ll get to that.

“Best” depends on who’s asking

A lot of neighborhood content starts with a ranking:

*The 10 Best Neighborhoods in City. *

But best for whom?

Someone who wants restaurants downstairs?

Someone with a child starting school?

Someone who works downtown five days a week?

Someone whose main requirement is that nobody plays bass through the wall at 1 a.m.?

Those people might love the same city and need completely different places to live.

That’s the question behind NabeWise: how can a comparison change when someone’s priorities change?

A person who cares about walking should see different options from a person who wants quieter streets. Schools should matter more when they’re relevant to the household. A commute should matter when there’s a workplace to commute to.

Otherwise, we’re all sharing one ranking and pretending we have the same life.

Which would certainly make product development easier. Slightly concerning for everyone else.

The interface starts with ordinary questions

On the city pages, you can adjust your budget and choose priorities such as walkability, quieter streets, or nightlife. There are also options for schools, work location, and commute time.

I wanted those choices to feel familiar.

“Do you want somewhere quiet?” is a question people can answer.

“Please configure your preferred residential amenity optimization profile” is a question that makes people close a tab.

For example, the Seattle page lets you explore different districts, change your preferences, and narrow the results.

You might begin by thinking you want the most walkable option. Then you realize that a shorter commute matters more. Or that you want a park nearby, but you don’t particularly need a street full of bars.

That’s useful progress. You’re working out your trade-offs before you start arguing with yourself over apartment photos.

A score needs an explanation

Putting a number on a card is easy.

Making that number useful is harder.

A score of 78 looks reassuring. It has the confidence of a student who actually studied for the exam.

But the obvious question is: *78 of what? *

NabeWise uses a weighted fit score based on several factors. The comparison stays within a city, and the weights change with the priorities you choose.

I’ve published the method behind the scores, because people should be able to inspect the assumptions.

The score is there to help someone narrow a search. It shouldn’t end the conversation.

If a result looks surprising, I’d rather someone understand why it appeared than assume a mysterious algorithm has discovered the objectively correct place to eat breakfast for the next three years.

Rent data comes with a surprisingly important footnote

The site uses Census ACS housing data.

That gives the comparison a consistent reference point, but survey median rent is different from the price of an apartment available today. It also combines different types and sizes of rented homes.

This matters in the interface.

A number beside a neighborhood name can easily look like a promise:

“You can rent here for this amount.”

That’s much stronger than what the data supports.

So the distinction needs to appear near the number, where someone is actually making a decision.

Nobody should have to open a methodology page after discovering that the beautiful apartment they want costs considerably more than the neighborhood median.

The apartment, of course, has “luxury” in its description.

It also has a dishwasher.

Apparently, we have reached the luxury dishwasher stage of housing.

Neighborhood names are their own engineering problem

People use neighborhood names casually.

Datasets need boundaries.

Those two things do not always cooperate.

Seattle is a good example. NabeWise uses the city’s Neighborhood Map Atlas districts, and some familiar names sit inside broader official districts.

Fremont, Wallingford, Green Lake, and Phinney Ridge are associated with North Central in the comparison. People can search familiar names, but the data still belongs to the larger unit.

That makes the geographic scale important.

A district-level comparison can help you decide where to investigate. It won’t tell you exactly what happens outside a specific building.

You can have useful data and still need to visit the street.

I find that distinction especially interesting from a development perspective: the search experience needs to speak the user’s language, while the calculations need a consistent geographic definition.

One wants “Fremont.”

The other wants a polygon.

Both have a point.

Walkability is more than a nearby coffee icon

NabeWise’s walkability index uses mapped amenities from OpenStreetMap, including groceries, cafés, parks, and transit stops.

That helps describe access to everyday places.

But there’s still a gap between a destination existing nearby and someone enjoying the walk there.

A map can show a grocery store close to an apartment. It can’t fully capture how you feel crossing a busy road with shopping bags in the rain.

And while coffee is important, being able to buy vegetables without organizing a small expedition is also a reasonable life goal.

The practical lesson for me is to keep the explanation close to the feature. A useful indicator becomes less useful when its label invites people to assume it measures more than it does.

Safety is the part that deserves careful wording

A safety score is easy to read as a verdict about a place.

That makes it one of the trickiest things to present.

Reported crime data has limitations. Visitor-heavy districts can have very different patterns from residential areas. Conditions can vary within a district, and an average won’t describe every block.

I want users to have access to relevant information without giving a number more authority than it deserves.

The wording around the score matters just as much as its color or position on a card.

This is one of those product decisions where “make it cleaner” can accidentally mean “remove the context people need.”

A very elegant mistake is still a mistake.

Design matters when people are comparing things

Alongside the data work, I’ve been refining the site’s colors, layout, and city sections.

Comparison pages can get crowded quickly. Each card wants to show a name, rent, scores, context, and a reason to click.

Give everything equal emphasis and the page starts feeling like every item is shouting its own résumé.

The hierarchy needs to help someone answer a few basic questions:

What area is this?
Why might it fit my priorities?
What would I be giving up?
Where can I read more?

I want people to be able to scan first and investigate second.

A nice skyline image can help someone recognize a city. Once they’re choosing between places to live, the information needs to do more of the work.

Nobody has ever made rent cheaper by adding a particularly attractive hero image.

I remain open to evidence.

Where I want to take it

The part I’m most interested in improving is the explanation behind a match.

A ranked list is useful. A clear reason is more useful:

“This option fits your budget and gives you better access to everyday errands, but your commute may be longer.”

That gives someone something to think about.

It also gives them something specific to disagree with, which is valuable feedback. “This feels wrong” becomes much easier to investigate when the assumptions are visible.

I’d also like to keep improving the descriptions, mobile experience, and clarity around data coverage. Some decisions need finer detail than a broad district can provide, and the site should make that obvious.

It’s live, and I’d like your honest reaction

You can try NabeWise here https://nabewise.com/

If you’ve ever chosen a neighborhood in a city, you didn’t know well, I’d love to hear what helped you decide—and what you only discovered after moving in.

If you build products with public data, I’d also be interested in how you handle the gap between a clean interface and messy underlying information.

That gap is a large part of this project.

The goal is to help someone get from “I have no idea where to start” to a shortlist they can actually investigate.

Preferably before their browser needs its own lease.

I Built a Small Site to Collect Startup Ideas from Real Product Launches

Auu Cooper — Tue, 06 Oct 2026 15:13:13 +0000

I spend way too much time looking at new products.

AI tools, tiny SaaS projects, developer tools, weird experiments on X, Product Hunt launches — basically anything that makes me think:

“That’s interesting. I wonder what else you could build from this idea?”

The problem was that I kept finding good products… and then forgetting about them.

A few days later, they were buried in my bookmarks or lost somewhere in my feed.

So I built a small site for myself:

startup-ideas.tech

The idea is pretty simple.

Instead of making another list like:

«100 Startup Ideas You Can Build in 2026»

I wanted to collect real products that people are actually launching.

Why I made it

I’ve read a lot of startup idea lists over the years.

They can be useful, but after a while they all start to sound the same:

Build a productivity app
Build an AI assistant
Build a tool for creators
Build SaaS for small businesses

The idea might be good, but it’s still very abstract.

I find it much easier to get inspired when I can actually see the product.

A 30-second demo can give me more ideas than a long article.

You start thinking:

What if this was made for a different industry?
What if it was simpler?
What if AI handled this part automatically?
What if this was built for mobile instead?
What would a cheaper version look like?

That’s the feeling I wanted the site to capture.

How it works

Right now, Startup-Ideas.tech is basically a curated collection of interesting products and launch demos.

You can browse different categories like AI, developer tools, productivity, design, and other areas.

I’m not trying to say:

“Copy this startup.”

Actually, I think that’s the wrong way to use it.

The interesting part is looking at a product and asking:

“What does this make possible?”

For example, maybe you see a new AI tool for designers.

The opportunity might not be to build another design tool.

Maybe the real idea is:

a version for real estate agents
a workflow for ecommerce stores
an API for developers
a simpler version for non-technical users

Sometimes the best startup idea is one small step away from something that already exists.

I’m building it as a small side project

This isn’t a huge startup.

At least not right now.

I’m treating it more like a small experiment.

I wanted to build something useful, put it online, see if anyone cares, and improve it from there.

There are still a lot of things I want to add.

Better search.

More categories.

Better filtering.

More context around each product.

Maybe trends showing what kinds of products people are launching more often.

But I’d rather launch something simple than spend months building the “perfect” version that nobody uses.

One thing I’m learning

While working on this, I realized that finding startup ideas isn’t really the hard part.

There are ideas everywhere.

The hard part is noticing patterns.

If you watch enough products being launched, you start seeing the same things appear again and again.

New technologies create new workflows.

New workflows create new problems.

And those problems create new products.

That’s probably what I find most interesting about building this site.

It’s becoming less of a “startup idea list” and more like a place to watch what builders are experimenting with.

If you want to check it out

The site is:

https://startup-ideas.tech/

It’s still early, so feedback is definitely welcome.

I’d especially love to know:

When you’re looking for your next project, where do you usually find startup ideas?

Product Hunt?

Reddit?

Hacker News?

Or just random problems in your own life?