When I started with Datespots.nyc, I knew what I was getting into. I had already lived in New York for almost nine years. I knew the boroughs, neighborhoods, parks, museums, cafés, and all the little corners that make the city what it is. Building the first version was difficult, but at least I understood the landscape.
Expanding beyond New York was a completely different challenge.
Almost all of the other 23 cities were unfamiliar territory. Every city has its own geography, culture, and way of organizing itself. Some are incredibly well structured, while others feel like the Wild West.
The hardest city by far was Mumbai.
Unlike cities with clearly defined boroughs or districts, Mumbai doesn’t have clean administrative or neighborhood boundaries that people actually use. Areas overlap, names change depending on who you ask, and Google Maps isn’t particularly helpful. If there’s one city on Datespots that I feel could still be significantly improved, it’s Mumbai.
But the problem wasn’t just geography.
After spending countless hours researching, I realized there simply aren’t as many unique date-worthy places as people might expect. I went scorched earth looking for them. I searched blogs, local recommendations, travel guides, Reddit threads, Google Maps, Instagram, and everything else I could find. Eventually you reach a point where there just isn’t much left to add without lowering the quality bar.
Most of the other cities, such as Amsterdam, Berlin, Madrid, Milan, Sydney, and Melbourne, were much easier from a mapping perspective. They have well-defined neighborhoods and districts, which makes organizing places far more straightforward.
That, however, is the easy part.
The real challenge is building a database that’s actually useful.
I wanted Datespots to stay inclusive. I didn’t want every recommendation to be another “Top 10 tourist attractions” list copied from the internet. The goal was to surface places that locals genuinely enjoy while still including iconic landmarks that deserve to be there.
Every city goes through the same process.
First comes gathering data from dozens of different sources. Then comes cleaning duplicates, verifying places, categorizing them correctly, checking neighborhoods, refining descriptions, finding city-specific experiences, and filling gaps until the collection feels complete.
On average, a single city takes about a full day of focused work.
It’s repetitive, tedious, and sometimes mind-numbing.
People often assume AI can do this automatically. It certainly helps, but it has limits. Large language models don’t actually know a city the way someone who lives there does. They hallucinate, repeat popular recommendations, miss hidden gems, and gradually lose consistency as the dataset grows. Human review is still essential.
And then there’s the Google Maps API bill.
Let’s just say it’s large enough that most people would be surprised.
So why were these particular cities chosen?
Because they share a few important characteristics. They’re generally safe, they have active social scenes, strong café and restaurant cultures, good nightlife, plenty of public spaces, and a younger population that enjoys going out and meeting people. They’re cities where dating isn’t just possible, it’s part of everyday life.
My Datespots platform isn’t trying to cover every city on Earth.
It’s trying to do a small number of cities well.