Live data from Hacker News

I need to find an apartment

mattrighetti.com

1–10 of 319 posts

Re: I need to find an apartment

#2
At least OP had the full address of those apartment on that website, so he could plot them on the map.

Here in Austria, most leeches, sorry, realtors, avoid putting the street in the ad to prevent you from trying to contact the property directly and bypassing their fees, so apartment hunting is even more of a headache.

Still, amazing work from OP.

Re: I need to find an apartment

#3
Interesting article: I did something similar but for Zürich (finding a low-price, but good apartment there is almost impossible).

I did follow a different approach though, and reverse engineered the APIs of the most common rental listing websites here.

I didn't publish the code (or better said, I didn't make it public), but it's similar to yours. Instead of SQLite I used a PostGis database where I stored the apartments with their point to point public transport distance from my office (pre-covid search). I did it in Rust.

Whilst the application is not available to the public (yet), all of the libraries I've created are now available.

I should have both a Rust and a Golang version for most / all of them.

On a side note, I did a similar project (scraper + visualizer / search) for finding a job for my girlfriend by scraping LinkedIn, Xing and a couple of other local job posting websites.

[0]: https://github.com/denysvitali/sbb-api-rs

[1]: https://github.com/denysvitali/homegate-rs

[2]: https://github.com/denysvitali/flatfox-rs

[3]: https://github.com/denysvitali/go-sbb-api

Re: I need to find an apartment

#4
I'm in the middle of an apartment search. I was reflecting the other day on how much searching for an apartment has changed over the last 10 years.

Conclusion: it hasn't changed much at all.

It's still necessary to search on a myriad of sites. Most of the time, by the time you inquire, a place is gone. Or the ad was misleading and that "$2000 2BR" is actually "$2000 for the cheapest Studio we have, but we also have 2BR for $3.5K".

Finding places that match my criteria still requires careful inspection of the description. Checking a filter box for "Parking" excludes places that only mention parking in the description.

With all of the modernization and streamlining that has happened in other parts of our lives, apartment hunting remains unstructured, messy, frustrating, stressful.

Aggregators like Padmapper helped for awhile, but these days, it seems like most listings are posted by scammy middlemen just trying to make a commission on the listing, and now the site actively makes the process even harder.

Is hunting for an apartment doomed to be this messy forever? It seems like an area that is ripe for disruption or a better option, but what would it take to actually solve this?

---

With all of that said, this post is exactly what I needed right now. I've been toying with a very similar idea to improve the efficiency of my search process and this is good inspiration.

Re: I need to find an apartment

#5
This is a great project and reminds me of a small project I put together when searching for an off-grid property. Price, property size, and distance to home were key constraints and using something like follium would have made it much nicer than the tabular format I used. Thanks for sharing!

Re: I need to find an apartment

#6

Interesting article: I did something similar but for Zürich (finding a low-price, but good apartment there is almost impossible). I did follow a different approach though, and reverse engineered the APIs of the most common rental listing websites here. I didn't publish the code (or better said, I didn't make it public), but it's similar to yours. Instead of SQLite I used a PostGis database where I stored the apartmen…

> I didn't publish the code (or better said, I didn't make it public), but it's similar to yours. Instead of SQLite I used a PostGis database where I stored the apartments with their point to point public transport distance from my office (pre-covid search). I did it in Rust.

How did you convert an address to long/lat for postgis?

Re: I need to find an apartment

#7
post #6

Interesting article: I did something similar but for Zürich (finding a low-price, but good apartment there is almost impossible). I did follow a different approach though, and reverse engineered the APIs of the most common rental listing websites here. I didn't publish the code (or better said, I didn't make it public), but it's similar to yours. Instead of SQLite I used a PostGis database where I stored the apartmen…

> I didn't publish the code (or better said, I didn't make it public), but it's similar to yours. Instead of SQLite I used a PostGis database where I stored the apartments with their point to point public transport distance from my office (pre-covid search). I did it in Rust. How did you convert an address to long/lat for postgis?

I was lucky, the data also included lat/lng and the address for most (if not all) the entries.

If I would have been confronted with such a situation though, I would have probably used some Geocoding API like Google Maps's geocoding [0] or OSM's Nominatim [1].

[0]: https://developers.google.com/maps/documentation/geocoding/o...

[1]: https://wiki.openstreetmap.org/wiki/Nominatim

Re: I need to find an apartment

#8
I appreciate the talent and ingenuity behind this apartment hunting project.

I just wonder if OP ever thought of contacting a broker? I would have done that before I wrote a single line of code.

Scraping all this data is really great, but a human broker is going to know the local area and the other human nuances of apartment searching in the local area.

Plus, brokers already have access to their own software that’s different than what’s on the public Internet, along with human connections. Why reinvent the wheel?

My last apartment came up as an option before it even hit any sort of public property advertisement platform, and that was thanks to our broker working directly with the landlord.

Re: I need to find an apartment

#9

At least OP had the full address of those apartment on that website, so he could plot them on the map. Here in Austria, most leeches, sorry, realtors, avoid putting the street in the ad to prevent you from trying to contact the property directly and bypassing their fees, so apartment hunting is even more of a headache. Still, amazing work from OP.

There are genuine privacy concerns at play here, for example there may be existing tenants/owners who don't want people randomly knocking on the door because someone saw an ad online. Or getting a bunch of junk mail because there is an extra data point about the property.

I guess the variety of concerns will vary from country to country depending on local laws / existing open data though.

Re: I need to find an apartment

#10
post #6

Interesting article: I did something similar but for Zürich (finding a low-price, but good apartment there is almost impossible). I did follow a different approach though, and reverse engineered the APIs of the most common rental listing websites here. I didn't publish the code (or better said, I didn't make it public), but it's similar to yours. Instead of SQLite I used a PostGis database where I stored the apartmen…

> I didn't publish the code (or better said, I didn't make it public), but it's similar to yours. Instead of SQLite I used a PostGis database where I stored the apartments with their point to point public transport distance from my office (pre-covid search). I did it in Rust. How did you convert an address to long/lat for postgis?

You could use nominatim [0]

[0]: https://nominatim.org

Post reply on HN