TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Using Deep Learning and Google Street View to Estimate Demographic Makeup of US

145 pointsby phodoabout 8 years ago

11 comments

d--babout 8 years ago
Dammit AI people, wake up! This is SO backwards!<p>AI-based profiling of anything is bound to filter off outliers. Censuses are made in the field for a reason. The goal is to <i>gather</i> data so that you can make statistical reasoning on it. Not the opposite!!!!<p>Here, they just gather some car data, and <i>infer</i> demographic data from it. And then what? We&#x27;ve just created a population which matches the car sample. We can&#x27;t draw ANY conclusion from that data, beyond the type of cars that are found in such and such neighborhoods.
评论 #13744036 未加载
评论 #13745757 未加载
评论 #13744699 未加载
评论 #13745744 未加载
评论 #13745868 未加载
stillsutabout 8 years ago
Everyone jumps to the dystopian scifi implication. But how about lets try thinking like real technologists?<p>Example use: I know of three small but vibrant towns in Massachusetts where my wife and I would like to open a Bed &#x27;n Breakfast. The problem is they&#x27;re too expensive. So I feed the street views of these downtowns into the AI, and filter by &lt;$250K home price. And it returns 20 beautifully &quot;quaint&quot; towns that I never could have found just by traditional census &#x2F; business database searching.
评论 #13746829 未加载
评论 #13746582 未加载
nlabout 8 years ago
This is a great piece of work.<p>I wonder if it would be possible to use a similar approach but on shop signs. Intuitively there should be some kind of relationship between the type of shops in different areas, but I guess it would be much less dense data than car type.<p>I guess the next step is to make a completely end-to-end version without the manual feature engineering.
评论 #13743982 未加载
Animatsabout 8 years ago
Or you could just order that data from Experian Automotive, which has vehicle registration data from all US states.
评论 #13743014 未加载
评论 #13742857 未加载
评论 #13745670 未加载
评论 #13746537 未加载
评论 #13749299 未加载
mempkoabout 8 years ago
Let&#x27;s ignore building racist AI for a sec. This is hugely flawed because most street view data is NOT updated on a yearly bases. Many images are years old. Doing yearly door to door will get fresher data.
Doctor_Feggabout 8 years ago
&gt; We focus on the motor vehicles encountered during this journey because over 90% of American households own a motor vehicle<p>This is a clever hack, but I sorely hope no actual policy decisions are made on the basis of a methodology that wilfully ignores the 10% who choose not to own, or can&#x27;t afford, a car.
评论 #13743887 未加载
anonuabout 8 years ago
From the paper on collecting Street View images: &quot;This was done via browser emulation and requires only the latitude and longitude of each point&quot;<p>My question is does the TOS for Google Maps or Street View allow for this?<p>I&#x27;m not trying to diminish the research - it is SUPER cool. I&#x27;m just thinking it would be good to have access to the dataset they captured if publicly available.
评论 #13746239 未加载
id122015about 8 years ago
I want people to use technology to find what those people who drive a car into pedestrians have in common. There must be something unique about them, the media says they all have mental health issues. Suddenly since 2016 people with metal issues got access to cars and trucks.
platitaabout 8 years ago
Extracting life quality related data from streetview does not seem to be a new thing: <a href="https:&#x2F;&#x2F;hackernoon.com&#x2F;machine-learning-our-cities-617ce005ba27#.eg22ymq75" rel="nofollow">https:&#x2F;&#x2F;hackernoon.com&#x2F;machine-learning-our-cities-617ce005b...</a>
misiti3780about 8 years ago
does anyone know how they download this data ?
h4nkosloabout 8 years ago
Weev is working on a similar project: <a href="https:&#x2F;&#x2F;youtu.be&#x2F;ZMptVkyZWE4" rel="nofollow">https:&#x2F;&#x2F;youtu.be&#x2F;ZMptVkyZWE4</a>