Train and Weather Tracker with Raspberry Pi and E-Ink

30 degrees and she's dressed in a puffer jacket? Wut?

I'd assume Fahrenheit, which is around -1 degrees Celsius.

At what point does one decides that a microcontroller doesn’t cut it and moves to a full OS system like rpi?

For example this could be an esp32 project

I made a board similar to this that is based on an ESP32 eink screen (the Inkplate, which I highly recommend). The easiest way to render a dashboard like this is with a website, versus making the ESP32 render custom components and call a bunch of APIs). There isn't enough power on the ESP32 to render a website, but there is enough to fetch and render an image, but something has to generate that image. On mine, I use Home Assistant and pre-render a custom dashboard into a PNG, which the Inkplate grabs and renders. This guy skipped all that by just having the Pi do all the work in the first place, which is a fine way to do things if you can get the display working how you like.

If you're producing in volume that's when the microcontroller and effort involved becomes worth the cost savings. For a static screen that just refreshes every 5 mins a tiny MCU can go a looong way. One could do this with 96k of RAM, the data is just text with a few sprites. As a bonus the whole device would consume a few mA on average.

I built something like this. I am amused that mine too was built for my wife and also was one of the few techie things I've ever done that received unqualified spousal approval.

Mine shows weather and tide times at my wife's rowing club. It has the form-factor of a large fridge magnet. This led to the need to be battery powered and therefore very frugal with power - so a RPi wasn't feasible. It uses an ESP32. The data itself is collated by an AWS lambda.

I love an eink project as much as the next person but the trains look like they’re so frequent that if you’re really after “calm” maybe you could just forget about the train times completely, walk when you’re ready and know there will be one along in no time…?

This comes from someone in the middle of the country where there’s no train (nearest an hour drive away) and the buses are twice daily. So, yeh, I’m a little bitter :-)

They’re probably not that frequent at all times, and in any case no point standing around a train station for ten minutes.

My nearby buses and trains are normally something like one every ten minutes, but I check times before leaving the house (albeit not on an eink thing).

Love those projects! Did something quite similar a few years ago, found it quite useful and there's something just nice about eink displays for showing info that normal screens don't have

https://lengrand.fr/complete-setup-epaper/

Very nice product that Inkplate. Seems like it's a sweetspot where I can think about the content and not worry too much about the hardware aspect.

Probably at about the point you look at implementing a network stack on a microcontroller that isn't the ESP32.

Yes, this could have been a ESP32 project, because that has a network stack included, but it doesn't have an LVDS LCD interface, so you're going to have to implement one or the other on any 'simple' microcontroller.

I have been using InkyPi for my own E-ink display. It recently got support for weather as well. I'm using a different display though, so your mileage may vary, but for me it works like a charm, would highly recommend. :)

https://github.com/fatihak/InkyPi

TRMNL community has published a lot of train/metro schedule recipe[1], that can be installed on any e-ink screen[2]

- [1] https://usetrmnl.com/recipes/

- [2] https://docs.usetrmnl.com/go/diy/byod

Disclosure: I work at TRMNL.

This sounds very cool. Did you document it anywhere?

Fun. I used an e-ink display for something similar (just the weather forecast) and opted to use the onboard ESP-32. I try to stay away from Raspberry Pi for simple things because it's total overkill to have an entire operating system in there.

https://blog.jgc.org/2023/04/a-personal-weather-picture-usin...

I was about to ask the same. 30° is a very uncanny number. Looks so much like summer temps in Europe

It's so frustrating to be walking into the train station, hear the train pulling in, rush to board and miss it by five seconds. Especially when you could have walked just a hair faster, or not stopped to look at that thing in the window for ten seconds.

I used an Inky wHAT display with an old Raspberry Pi B+ to build a weather display that shows live weather pulled from my Ambient Weather station. My project uses Inky's python library. Since the Ambient API doesn't provide forecasts, I pull a daily forecast from the US Weather Service. And I grab the moon phase from Visual Crossing. https://github.com/skypanther/inkywhat

The Waveshare e-ink display he's using is $172.99.[0] The one that jgrahamc used in a sibling comment is about $200.[1]

It's a shame that it's not easier to pull the e-ink displays out of used e-readers because you can buy old Kindles on eBay for about $20-40. I know you can jailbreak them and achieve something similar to OP, but it'd be neat if you could just drop the vendor's e-reader firmware entirely and run your own software.

[0] https://www.waveshare.com/product/displays/e-paper/9.7inch-e...

[1] https://soldered.com/product/inkplate-10-9-7-e-paper-board-c...

Thanks for sharing, just made a pre-order, exited to have such a device, but as a young parent of two, I do not have the time to tinker such a costly project.

a few of our users open sourced their own public transport (e-ink) plugins here, for major cities like NYC, Chicago, London, Zurich, etc: https://usetrmnl.com/recipes

I recently got it. Device and firmware is a good value. I wish device would have better plastic quality but for something that hangs on the wall it doesn’t really matter.

It was very easy to make my own server with reactjs rendering for it. Now, adding new views is super easy and customizable.

I actually have a project like this, it's an alarm clock though. The idea was to not have something additional glowing at night and could pull my alarms from an online calendar. I've been procrastinating rewriting the code for an esp32 though since I think using my raspberry pi for an alarm clock is kind of a waste.

I'm wondering if it's possible to track ANY operational train, not just passenger trains. I'm a railway fan but I'm only interested in fright trains. I guess that piece of information is somewhere within the railway system but not public. And it's probably illegal to get a hold of such information.

Edit Found these: https://asm.transitdocs.com/map

https://railrat.net/ (probably just passenger trains)

I recently saw that my local supermarket had replaced most of the paper price tags with eink displays. I guess they can be updated dynamically from head office. Seems like you can buy them now on ebay for not much per unit.

https://tinyurl.com/28v4f583

I wonder what their battery life is, and which wireless protocol they use.

This is one of those projects every nerd loves to build :D

My first iteration back in lockdown [0] focused mainly on rain (I'm from the UK), but later versions used a colour e-ink, added a camera to change the background with a selfie and added train times from my local station, along with upcoming birthdays. Reminds me I need to try and write an update at some point!

[0] https://jamesjarvis.io/posts/2020/06/umbrella-e-ink-weather-...

I have an old kindle I was keeping in a draw as I planned to do something similar as the OP when I had time.

I've done no research yet though, what's the problem of reusing a kindle screen? It's not like old displays where you can plug a ribbon into something else?

3-5 years typical, when I was looking at the newer ones at an auction site to play with most were IR with some bluetooth etc..

The link you provided appear to probably be IR based, but it was one of the brands that was a dead end for official docs when I was looking at them.

The newest ones seem to be using Periodic Advertising with Responses (PAwR) that was introduced in Bluetooth 5.4, but I couldn't find those type for a price I was willing to pay.

Nordic Semiconductors has an Apache licensed SDK and RToS, I didn't find the device and time to play with it but bookmarked it.

https://www.nordicsemi.com/Products/Development-software/nRF...

Maybe you can, but I haven't seen anyone do that. It is tricky because the cases are apparently glued shut, so you need to do some delicate surgery to access the screen without breaking the device.[0, 1]

[0] https://blog.lidskialf.net/2021/02/08/turning-an-old-kindle-...

[1] https://www.instructables.com/Amazon-Kindle-e-ink-Screen-Tra...

No connection to this project, but I also have a couple sitting around for "some day" The biggest issue is that most of the adapter boards that are often available are not that great, so having a board built from EDPiy is an option.

https://github.com/vroland/epdiy

For the kindles that have wifi, custom firmware is probably easier?

TRMNL https://usetrmnl.com/ is an e-ink display that offers many plugins, called ‘recipes’ — and you can build your own. price is ~ $150 and you can opt-in for a ‘jailbreak’ version, too. all open-source.

That particular display that they use is around $40, and then there's a $3 ESP32.

I bought one because I think the software and the ecosystem is worth the price, but it's not terribly hard to make one of these for much less than $150.

I did something similar for my daily calendar. I don't have a writeup of the actual device, but here's the firmware that generates the calendar image:

https://github.com/skorokithakis/calumny

It's a really fun project, but I wish these Waveshare displays had more open firmware. I didn't really manage to get the ESP32 to talk to the display (or even to connect to it, as I don't have the right adapter), so I used their sample code with its weird API and I'm sending an image straight to the ESP32, rather than have the ESP32 pull it from an HTTP server, as I wanted.

I don't trust the founder, and I don't want to do business with him.

His previous company was called fomo, which added popups to ecommerce stores to manipulate users into buying by claiming that others had recently bought the same product. I haven't dug into it to prove that the data was fake, but it's highly likely it's fake.

He sells marketing courses where he boasts about techniques like exploiting a website vulnerability to spam 20k people who never engaged with him at all.[0] He also proudly shares how he lied to 78k inactive users to trick them into responding to an email.[1]

[0] https://www.ryanckulp.com/targeted-outreach-email-with-20-li...

[1] https://www.ryanckulp.com/birthday-reactivation-hack/

jumping in to defend myself (thanks f5bot). you are not a serious person.

1. i'm very proud of Fomo. for 6 years we fought against fake data. i said "no" to many lucrative opportunities that were not ethical. our vision was "to given entrepreneurs the credit they deserve," and our mission was "to help consumers make better buying decisions online." your comment is libelous. [1, 2]

2. as of yesterday, all 11 of my courses are free. i've helped over 10,000 people learn to code and learn to sell. [3]

3. on the side i do a bit of pen-testing. i help stop* scammers, and i help non technical folks protect their brand. that's why Bubble added me to their home pages and dozens of cash bounties have been sent my way. [4, 5, 6]

[1] https://blog.fomo.com/having-a-vision-works/ [2] https://blog.fomo.com/introducing-honest-marketer/ [3] https://x.com/ryanckulp/status/1907099032616378815 [4] https://www.ryanckulp.com/hustlewing-is-a-scam-so-i-hacked-i... [5] https://bubble.io/security_acknowledgements [6] https://bubblesecurity.gumroad.com/l/unhackable-101

thanks for the support! re DIY, we encourage this: https://docs.usetrmnl.com/go/diy/byod

Nothing you said refutes any of my criticism.

Since you're accusing me of libel, I just checked the fomo website, and I can see that right now, it's presenting fake data. It shows a live updating counter of the number of fomo's impressions and clicks, but these numbers are updating every 200ms based on client-side JS, not real data. From the HTML source:

    setInterval(function() {
      var serverTimestamp = 1743654982;

      // impressions
      var imprCountWhenChecked = 19546526366;
      var imprFrequence = 80;
      var freshImprCount = Math.ceil(imprCountWhenChecked + (Date.now()/1000 - serverTimestamp) * imprFrequence);

      document.getElementById('statisticsTotalImprCount').innerHTML = freshImprCount.toLocaleString('en-US');

      // clicks
      var clickCountWhenChecked = 43316684;
      var clickFrequence = 0.15;
      var freshClickCount = Math.ceil(clickCountWhenChecked + (Date.now()/1000 - serverTimestamp) * clickFrequence);

      document.getElementById('statisticsTotalClickCount').innerHTML = freshClickCount.toLocaleString('en-US');
    }, 200);

I don't have a clean copy of the HTML source when you were running fomo, but I can see on an archive.org snapshot[0] that the customer count increases by exactly 7 every few hundred milliseconds, so it looks like this functionality was introduced while you were running fomo.

If the site fakes that data and you've publicly admitted to lying to your users in other cases, it increases the odds that other data you presented was inauthentic.

[0] https://web.archive.org/web/20181101030710/https://fomo.com/

ok back up - you claimed we used fake data. obviously you are referring to "John bought X, Y mins ago" notifications. the actual product.

for our homepage counter, this is not customers. this is notification count. with thousands of connected websites, each showing multiple notifications per page load, we quick racked up billions of notifications shown.

running `select count(*) from events` on every landing page load is untenable. instead, every N period we'd check on our avg daily notification count, then update the landing page JS to give users a sense of scale. this number was perpetually under-reporting our actual metrics.

for metrics that actually matter, like signups, customers etc, we created an Open API and read from those metrics on various landing pages: https://fomo.com/open https://docs.fomo.com/reference/fomo-open

and for the core product notifications, "John bought X, Y mins ago" we only allowed data from native integrations. if we saw someone pinging webhooks to us with fake data, we deleted their account. if someone asked for a CSV upload feature, we deleted their account. i cannot say the same for our competitors, who welcomed these use cases with open arms.

bottom line - you're a hater. you don't like Fomo (fine), so you're trying to sabotage the sale of another, completely unrelated product. think about where this bitterness comes from and do better next time.