A “Rat” Every 14 Seconds: Analyzing NYC 311 Calls | Hacker News Reader

A “Rat” Every 14 Seconds: Analyzing NYC 311 Calls

omnisci.com

60 pointsjonbaer30 comments

I made a front end to search for individual buildings' rat inspection records. Working on predicting risk too - needs some tuning, the "rat model" is less accurate than the one for code violations, which you can try now.

https://beta.getaugrented.com

Now for LA

I love it! Will try to make a request for the same type of data from Oslo municipality (Norway). We have a law where they have to hand over the data. Think we have a lot to learn from NYC.

> Will try to make a request for the same type of data from Oslo municipality (Norway). We have a law where they have to hand over the data.

Is asking them to spend time getting this data for you honestly a good use of precious public money?

Transparency is always a good use of resources.

Always? No exceptions? You can't conceive of any situation whatsoever that spending time responding to a frivolous freedom of information request is a better use of a civil servant's time over meeting an actual immediate need of a member of the public?

Freedom of information is not frivolous.

I can see you editing your comment out from underneath my response, removing your abuse of public servants.

No individual freedom of information request can be frivolous? I can't understand how someone can think it can never be a waste of finite time and resources to respond to a request.

Transparency does not inherently require a request.

It’s the government’s fault that the data isn’t easily accessed by citizens. No employee time would be necessary if data was accessible by everyone without having to go through a worker. I know this firsthand because there are places in the US that you have to submit requests and deal with government lawyers and slothful employees, while the exact same type of data can be directly downloaded from another municipality without any barriers or assistance.

Is it acceptable to have unmarked advertisements on HN?

I like questions that show how useful it is to have only one thing you're optimizing for. They're easy to answer and the answers are sometimes surprising.

On HN, the only thing we're optimizing for intellectual curiosity, so to the degree that a post gratifies intellectual interest, it's fine. If the 30 most interesting things on HN happened to be ads, then the best front page that day would be all ads. Of course that won't likely happen before the universe ends, but still, it follows!

The problem with content marketing is that it tends to be lame. If someone produces a genuinely interesting article in the genre of "we used our product to do X", that's not only ok, it's interesting in another way too: "we managed to make content marketing not be boring".

One thing that's not ok, of course, is for people to get others to upvote their content marketing in order to promote it. But that applies to every submission. And the votes on this story look clean.

As a good example, I remember a whole string of great Priceonomics articles showing up on HN—well-written, interesting investigations into cool topics. They might have been part of a content marketing campaign, but they were also consistently in the top 10% of articles I saw on the front page.

And they were mostly all plagiarized from other sources.

dang and the other moderators are happy to update submissions to original sources when folks leave a comment and the original source better satisfies HN's goals.

But I do acknowledge that's a relatively recent change (last few years).

That's a severe charge. Please don't do that in an unsubstantive comment.

There's no guideline against it as far as I know.

Anyone talking about what they do is an 'advertisement.' I don't know what good marking them would do - you should be as skeptical about anything anyone tells you whether it's an advert or not.

I strongly support "adverts" like this and I wish that all companies spent a far higher portion of their marketing budget on creating content like this rather than on targeted advertising.

To me, it seems like a win for everyone.

- I hear about the company while learning something interesting. I'll probably remember their brand.

- The company has people actively searching for and promoting their article (e.g. on HN) instead of trying to bypass or block it with adblockers etc. And this article might bring in leads and sales for them years in future, instead of only while they keep throwing money into adspend.

- A writer who is interested in technology and writing gets paid, instead of that money going to Google / some ad agency who randomly puts levers on and off hoping to make the conversion rate go up.

- An editor, designer, proofreader, technical checker etc etc all get involved (as opposed to a one-person blog piece with a similar goal of "look at this cool thing I made") -- also people I would like to see supported more over ad agencies.

I think (hope) that "marketing as education" -- e.g. actually high quality content that also promotes a specific brand instead of the cheap "content marketing" that so many people are trying now -- will be the future of marketing once people realise they don't want to be tracked so invasively, adblockers get better, laws tighten, etc.

I'm just glad NYC 311 turned out to be useful for someone.

I actually had an experience where 311 was extremely useful! A couple of years back, my grandmother visited New York, and when we got back to her hotel after dinner, she realized she had forgotten get pocketbook in the cab. Since I had paid for it with my credit card, I was able to get the cab number from my bank's app, but I was stuck without a way to get the cab driver's contact information. I was fairly new to New York at the time, so I didn't actually know about 311, but the hotel concierge recommended I try it, and I was able to get the phone number of the garage for the cab. I called them and let them know what happened, and they called the driver and gave him my phone number, and he drove back to return the pocketbook. I'm not sure if there were any other ways of looking up the garage from the cab number (I spent a good amount of time googling without any luck), but 311 was able to do exactly that, which was pretty astonishing to us.

That’s a good story, and I’m glad again, thanks.

cached: https://webcache.googleusercontent.com/search?q=cache:tPeRrJ...

Are they down? I'm unable to open the link.

Wayback: https://web.archive.org/web/20190614191130/https://www.omnis...

More stable mirror: https://web.archive.org/web/20190614191130/https://www.omnis...

And for basically an advertisement for the Omnisci product written by a "developer advocate" (which we used to call a solutions architect at Sun but was really just a sales person).

TL;DR version - If you can import your data into our product [here let us show you how] you can visualize many different things [here let us show you how] and using your new insights take actionable steps [here let us show you how.]

I keep hoping for the equivalent of the Jepson call me maybe series but for data visualization packages. That and perhaps a 'youtube of data' which are data sets loaded into a common visualization framework so that you can browse different data sets like you do videos on Youtube.

I, too, would like to see the OmniSci Community team make a Call Me Maybe lip sync. I’ll post the suggestion in their Slack channel (disclosure: I’m an engineer at OmniSci).

A challenge for us with a “YouTube of Data” is that our backend requires GPU instances for the dataviz that we focus on. Scaling that to YouTube heights would take a really well concerted effort, more than a fun side project. We’ve talked about the idea a lot though. Our backend (OmniSciDB) is open sourced so anyone else could do it with their own resources and a custom dataviz client.

I know I’m sounding like a commercial, but I thought you might actually like an answer. That’s basically how we play with the app internally. I like using our Pokémon data set for debugging whenever possible :-)

Our paying customers are big enterprises right now who are well supported by the kind of sales team you’d expect. Our Community team is totally separate from sales and they genuinely likes helping people use our open source offering, hence the demo. We’re pretty happy for Sean that his post made HN.

But yeah, we’re still a business. But we’re kind of a family too. And data junkies.

Other threads on NYC 311 calls:

2018: https://news.ycombinator.com/item?id=18316622

2010: https://news.ycombinator.com/item?id=1870111

Unrelated, but one of my favorite things about HN is going back and reading old threads like this. It's a fascinating anthropology of how we view tech over time.

This is awesome. Love seeing people dive into publicly accessible data, make sense of it, and highlight features which really stand out.

Good use of a data set and illustrates what's possible.

Reply on news.ycombinator.com