I wanted to draw an analogy to wardriving: instead of collecting my neighbors' Wi-Fi SSIDs, I collected information about their houses' characteristics and quality. Pehaps it is a stretch, but the original idea was actually to drive with a camera on the roof of the car, but it proved to be a costly setup, and Street View was more "scalable."
For the next part, I will actually need to do it: Drive around, perhaps recording, and then extracting "for sale" signs. I have no idea how I am going to make it cheaper, maybe using simple text detection instead of an AI will suffice.
I was honestly looking for / intrigued by (the thought of) e.g. a correlation between Wifi 6E access point density vs older ones, and quality of living / neighbourhood.
sounds like OP TFA just drove around! sounds like how everyone and my grandma finds their house. it’s like tptacek says: ‘talk to me when you get deep in 801.11!’ -tptacek via armondo
This city doesn't look like much work to drive. One evening of work to scope out every nook and cranny is infinitely better than what's possible for most American cities!
It is surprising the lengths some people will go to trying to code their way out of a problem rather than interact with the nearby and available human beings who already have the answers*.
* Yes, I too have been guilty of this in the past >..<
Exactly. All the person had to do was strike up some conversations in restaurants, the laundromat, whatever. "I love a neighborhood full of greenery, where should I look?" or "I love to eat out, what neighborhoods have really great restaurants?"
An AI will never be able to tell you that the "nice looking" neighborhood is full of snooty wealthy people who keep to themselves and rat you out if you do anything not up to their standards - whereas in the kinda rundown looking neighborhood everyone knows everyone, there are fun block parties on holidays, and if you need anything, people will do whatever they can to help.
Police logs don't work because police love to overpolice poor/minority neighborhoods.
Property values don't work because property values have little to do with how strong a neighborhood's community is.
Another data point is property tax records, some of which will list building "condition" on a scale of 1-10, or look at assessed property values over time.
It is nice to see millennials re-discovering how pretty much everything was done before the explosion of the internet in the mid-90s.
Historically, what this article describes, was simply called "finding somewhere to live". There was no necessity to invoke warfare when looking for a house. One simply drove around the neighborhood looking for for-sale or for-rent signs.
Amazing that now just doing something in the real world is cause for excitement! But I'm sincerely happy to see a re-emergence of the practice.
BTW, I hope the author appreciates living in a place where "there are almost no real estate agents" along with the other described traits. Escaping the commodification of everything is more difficult every year, and you have to get pretty far from the epicenters to find it at all.
While a beautiful execution of an idea, he should've just taken a few days to drive around the neighborhoods and talk to people on the street and in the bars and restaurants.
There are still a few tasks where there is no substitute for ground-truth real-time cultural immersion.
This. By visiting venues such as supermarkets, you get a good idea of the vibe and demographics of the area. Of cource, you'll want to assess the property itself as well :)
This was fun to read. 8 years ago I used the google maps api to group apartments in Seattle that had fiber internet by neighborhood (http://blog.alexchantavy.com/articles/using-the-google-geoco... your blog is a much much more modern spin on house hunting. And yeah others are saying you could’ve just walked around and talked to people - and I’m sure you will - but this is still a useful way of showing how AI can be used to classify things for hackers in daily life.
Interesting idea. I think it becomes far more interesting if you can come up with analysis that not only provides homes for sale but homes that may be coming up for sale. Which neighborhoods have listings for furniture on Facebook Marketplace? Which homes appear to no longer have anyone living there. I look at a home I own that will be sold soon. No internet is setup there. Maybe you Wardrive, by the literal definition, and find homes that don't seem to have internet as an indicator that someone may be interested in moving/selling. The possibilities are endless.
The motivation behind this project was the fact that the city has limited online property listings. Hence, wardriving. I guess that the availability of other online data is then also limited.
Yeah that is a good idea, for internet I was surprised that the city has lot of fiber companies so lots of wifi all around, I pay just $12 for 800mb/s which is cheaper than metropolitan cities (and that's the "expensive plan").
Facebook idea is awesome, I was trying to find posting on "craigslist" alternatives but got nothing, but since the town skews older, they are all must be on Facebook!
The headline made me think someone was wardriving around looking for spaces with no wifi signal to find places to occupy/squat in (legally or illegally, depending on where in the world this would be), but it seems more like scanning street view for interesting places.
https://en.m.wikipedia.org/wiki/Wardriving