a few days ago, while browsing the fediverse (decentralized social media), i stumbled upon 2 posts about a geolocation challenge. those were writeups by 2 seperate beings: one by annie, and one by ryan. they talked of an image that had been sent into a discord server by maia, which they needed to locate (similarly to GeoGuessr).
both posts were encouraging others to try it, and included some crucial information. however, i wasn’t sure about the exact rules, since in GeoGuessr, the #1 (unwritten) rule is that you can’t Google (tied with not being able to use scripts and having to curse Google for terrible movement in Street View). so i messaged annie for clarification, and it turned out i could use anything. “ohhh this will be interesting and fresh”, i thought. i also decided to think out loud in those messages, with the potential for hints, but also for fun (this will help me a lot with recalling how it went tho!). spoilers ahead!
there were 2 texts i noticed first: one on the pole, one on the train. i decided to go with googling the train’s number at first, cause utility poles rarely have databases surrounding them. with trains, tho? there have to be a few transit nerds in every country. i should know, i interact mostly in queer and nd circles. after googling “swiss rail rbde”, the first train model that came up was the SBB-CFF-FFS RBDe 560, and i definitely thought it could be a 0 (note: while i didn’t know this, it definitely curved on the top left - so my brain could certainly make that autocompletion). i think i also googled something similar with “red dots” in the search phrase, which did yield a result with a rail picture sharing website having a photo of an RBDe 567, operated by TPF. they also mentioned what line they were on, so i looked at it on Google Maps (not sure if Street View), and i found out that the red dots probably did mean TPF. yay, a clue!
their wiki page said that their main line runs between Bern and Lausenne. okay, we have some stations to look at now! the image’s file name was PXL_20230917_012607611. fine. surely, swiss rails run on time, so i’ll just see if i can find a database of past trains. i look at the swiss national rail’s db website - oops, i’m doing it 2 days afterwards and the db only has the previous day’s runs. well, fuck. (why don’t i think it runs at the same time all days? i don’t not think that, that completely slipped my mind by this point). but i’m in luck. turns out there’s a comment section. it runs on fucking disqus.. great lmao. but one of the comments has a link that contains past several days’ databases. let’s fucking go! so i click download and i have a high speed connection, so this will be down in 2 seco- oops it’s 300MB. don’t get me wrong, it downloaded quick. but. it was a 300MB csv file. fine. i’ll open you in libreoffice. what could possibly go wrong?
well, it turns out libreoffice couldn’t display all lines in the first place. as it happens, Switzerland, the country partially known for having a lot of rails… has a lot of trains as well. damn. sorting them would be an even harder task. okay, fine, let’s open it in a text editor! it stops working properly as soon as i hit ctrl + f. i also go onto TPF’s website. and i suffer. is the website terrible? well, yes but not for design reasons. it’s because it only has 2 of my most hated languages available: french and german. i did notice that they mentioned freiburg a lot, but decided not to pursue that. well, there was only 1 solution i could think of. since the timestamp was in the filename, and swiss trains mostly run on time, plus knowing it’s TPF, which is just one company, maybe i can filter those out from the database with a short python script!
and i got 4 results! none of them that close to 01:26 (i was filtering for anything “01:”)… surely it’s an active train, the lights inside are bright? yes. i felt like i hit a wall, to which annie said it may be worth exploring some of my leads further. that was also kind of a hint that i’d been on the right track at some point! this was about 50 minutes into trying. i was looking at the TPF website, for schedules. i did find some! and some of them ran in close to 01:26, but all before then by at least about 10 minutes. still, maybe it had just stopped at the time of image? my messages in discord were the following ("|" seperates messages): “their last trains stop around 1, mostly | many freiburg references | one of them is in bulle | bulle has painfully few pics on gmaps | i don’t see it being freiburg, but there’s a good chance i wouldn’t actually realize lmao”. i was trying to check them using images and Street View, but it was hard like that. so after a suggestion, i tried 3d Google Maps (which is a lot like Google Earth). i’ve sent this screenshot for demonstrational purposes of me using it (may have been Google Earth, i also opened that for a bit):
i still couldn’t find it, however. i asked if i was on the right track, and got a yes. i did try making more observations on the image 1h10min in. i saw there were multiple pairs of rails, maybe 3. the roof being dark was very annoying, however. i just kept looking, i got the first actual hint i got at about 1h45min in. it was about what makes the building kind of unique (so that i could find it). but oh dear, it did not help! as it turned out, it wasn’t my building comparison skills that were lacking. so i decided to ask to make sure it was in fribourg (bilingual city, naming very much confused me by then), and it was! i was right about that. i kind of gave up around this point. i’d been looking at that railway station in 3d for more than 30 minutes, so i wasn’t gonna find it. so what was the problem? i was looking directly by the railway. it wasn’t. the top left building on the above screenshot i’d made (50 minutes earlier by that point) happened to also be the building that was in the background. the perspective i thought of the original image was very different than what it ended up being.
i suppose i ultimately did find the city. and i did even send a screenshot of the building in the background. i just happened to do so while being unaware of it the whole time. and the challenge wasn’t to send a screenshot of the building, it was to find the location. so it was a semi-success! but the punchline of the story is that the time stamp was in UTC the whole time, while Switzerland is in UTC + 2 during DST. so all those schedules? they didn’t matter. that train was asleep the whole time. it was supposed to, anyway. maybe this train is a night owl…
