Love this! I did a machine learning class in my undergrad, largely before the LLM+genAI boom. Mostly classifiers -- similar stuff to this too. Thought this stuff was fantastic when I got something working, and surprisingly easy to get started in Python once you got the basics down. Kind of miss it. I had these kinds of cool ideas but never had the time then, or now, to re-learn and do anything interesting. Who knows, maybe one day I'll get back into it, things like this inspire me to.
In a previous office location, there was a coffee shop a block down that quickly became my favorite. It was a few years before I found it though since it didn't show on google maps unless you zoomed ALL the way in on that exact spot.
I'd love to see a similar map for Seattle if anyone makes one.
I use Yelp’s “Hot and New” list to break away from the natural advantage of older restaurants. I focus on restaurants that started in the last few months with fewer than 20 reviews. New restaurants tend to be more interesting and they try harder because every review makes/breaks them. It’s fun and I get to enjoy restaurants for a year or two before they take off.
I was positively surprised by the quality of this post. There are many posts about one-off data experiments on real-world data, it is the best I have seen in a while. The experiment seems to be done well (I am no expert) with interesting technical remark, the topic is interesting, the writing is very good.
Also the biases really show up when you travel to foreign countries. I’m currently in Vietnam eating based on recommendations from friends I trust, recommendations from guides and staff, recomendaciones from our travel guide book, and Google reviews.
We use Google reviews to remove obviously touristy places and pick a place if it has at least two or three dishes we want to try based on photos. Most places Google recommends are tourist or expat friendly and that’s not why we travel around the world to eat. Often international foods, mostly euro-centric. This is especially sad cause there are many other Asian foods in Vietnam like Japanese, Thai, Cambodian, Indian, or Chinese that can be good but just aren’t desired by European travelers. Finally, “sponsored” places get a negative. Often they’re chains or more about presentation than good food. I have yet to be proven wrong.
So far, our masseuse gave the best recommendation. Then our family friend is a good second place. Finally Google Maps but limited to just a few blocks and I clicked around the map looking at photos.
Love this! I did a machine learning class in my undergrad, largely before the LLM+genAI boom. Mostly classifiers -- similar stuff to this too. Thought this stuff was fantastic when I got something working, and surprisingly easy to get started in Python once you got the basics down. Kind of miss it. I had these kinds of cool ideas but never had the time then, or now, to re-learn and do anything interesting. Who knows, maybe one day I'll get back into it, things like this inspire me to.
In a previous office location, there was a coffee shop a block down that quickly became my favorite. It was a few years before I found it though since it didn't show on google maps unless you zoomed ALL the way in on that exact spot.
I'd love to see a similar map for Seattle if anyone makes one.
That's some good food for thought
I use Yelp’s “Hot and New” list to break away from the natural advantage of older restaurants. I focus on restaurants that started in the last few months with fewer than 20 reviews. New restaurants tend to be more interesting and they try harder because every review makes/breaks them. It’s fun and I get to enjoy restaurants for a year or two before they take off.
Excellent analysis and writeup.
I was positively surprised by the quality of this post. There are many posts about one-off data experiments on real-world data, it is the best I have seen in a while. The experiment seems to be done well (I am no expert) with interesting technical remark, the topic is interesting, the writing is very good.
Also the biases really show up when you travel to foreign countries. I’m currently in Vietnam eating based on recommendations from friends I trust, recommendations from guides and staff, recomendaciones from our travel guide book, and Google reviews.
We use Google reviews to remove obviously touristy places and pick a place if it has at least two or three dishes we want to try based on photos. Most places Google recommends are tourist or expat friendly and that’s not why we travel around the world to eat. Often international foods, mostly euro-centric. This is especially sad cause there are many other Asian foods in Vietnam like Japanese, Thai, Cambodian, Indian, or Chinese that can be good but just aren’t desired by European travelers. Finally, “sponsored” places get a negative. Often they’re chains or more about presentation than good food. I have yet to be proven wrong.
So far, our masseuse gave the best recommendation. Then our family friend is a good second place. Finally Google Maps but limited to just a few blocks and I clicked around the map looking at photos.
I’d love to see this data in cities near me. :)