Under construction: cheapbots, tracery, other stuff. Adding information as I port things to mastodon, so this is going to take a while…
Most of the Mastodon bots are on the botsin.space instance, thanks botsin.space
!!
Where I’ve needed lists of different types of thing, Darius Kazemi’s excellent corpora
.
cheapbotsdonequick (and it’s mastodon equivalent cheapbotstootsweet) are extremely good ways to make bots that require so much less faff than writing them in R. So I use it a lot. It’s great.
I’ve included the tracry sources here in case there’s something useful for others to see in my totally derranged JSON.
@transect575
twitter // mastodon // twitter tracery source // mastodon tracery source
When I worked for REDACTED GOVERNMENT AGENCY I saw a lot of aerial survey data from people looking for whales in the sea. Usually this includes useful information like the number of animals, whether they are juvenile or adult and weather conditions. They also usually included a comment field. So I scraped that out and this bot makes haikus from it.
There was some faff in working out the syllable counts using the qdap
R package. There was also a bit of faff in anonymizing the data, hence no public source to this (private source here to help my memory).
@birdcolourbot
twitter // mastodon // tracery source
The Cornell Lab of Ornithology’s Merlin app had information about the colours of birds. In the app this helps users identify birds by their colours. But instead we print those colours out for fun. This bot does that.
On a technical level, this consisted of futzing around with the Merlin App .apk
file to extract the JSON which has all this data in it. At some point I should do this with a newer (i.e., post-2016?) version of the app, as there’s now much better coverage of the app (when I first did this only North America was covered).
Somehow this bot got some interest in the New York Times (!)
@clapping_bot
/@clapping_music
twitter // mastodon // twitter tracery source // mastodon tracery source
Do you like Steve Reich’s Clapping Music? If so then this bot is for you. It toots emojis showing the parts of Clapping Music.
👏👏👏 — 👏👏 — 👏 — 👏👏 —
@RverbsR
twitter // mastodon // twitter tracery source // mastodon tracery source // R source to generate tracery
At some point I got a brain worm from the phrase “verb that noun”. Then I got thinking that R functions are kind of verbs, and data types are kind of nouns. This bot just mashes-up these two based on a list of each.
My setup is on a Mythic Beasts (who are awesome by the way) virtual server. Each bot is basically just an R script, which gets run by cron
whenever I feel like is appropriate. Internally the bots use twurl
to post to Twitter (there’s a bit of fiddling in ensuring that the right twurl
profile gets used with the right script). For Mastodon, I use the excellent rtoot
package.
@goodsdmbot
This is really a Work Bot. In species distribution modelling, it’s not uncommon to see people making some truly bad maps of where they think animals are and then overinterpreting them. This bot makes fun of this practice by randomly picking climate variables, randomly generating coefficients and then plotting the resulting addition in a nice colourscheme. It also creates a fake species name. Next stop, Nature.
So there’s a bit in Jurassic Park (1996) where Lex needs to close some doors (or something) and sits infront of the computer and realises that it’s a Unix system, and she knows this. It’s a great moment.
So, I wrote a bot that imagines universes where the computer was running different operating systems. Like Palm OS or Windows.
mastodon // github (bot) // github (images)
This is a really simple one. It just posts screens from Adam Curtis movies, specifically those where there is some text on the screen.
mgcv
(@mgcv_updates
/@mgcv_changelog
)This is a simple little bot that just tweets when there are changes to the Changelog
file for the R package mgcv
.
These bots won’t get ported to Mastodon or fixed on twitter if they break. They were usually the result of a particularly stupid news story that probably isn’t that funny any more.
@listbot3000
This bot mashes-up two Wikipedia lists categories imagining what it would be like if there was a non-empty intersection of the two sets.
(I do have a plan to write another bot at some point which checks for the intersection and reports it, but I’ve not got around to doing that yet.)
@BrexitNewsBot
After the incredibly stupid move of voting to leave the EU, the tabloid/tabloid-adjacent-right-wing press ran a lot of news stories saying how particular niche parts of the population would be better-off post-Brexit. I thought it would be funny to make a bot that comes up with random scenarios, in the form:
A #politicalideology# think tank published a report that #groupofpeople# will be #percentageorpounds# #betterworse# off per #absoluteunit# as a result of #brexitstatus# Brexit.
@excellentwmoney
Based on a completely infuriating article on CNBC about how “young people” budget.
@leastbadhouse
During 2020 when we were all in lockdowns, there was a popular meme on twitter asking people to select between “quarantine houses” that they’d prefer to be stuck in. This bot generates such houses with the help of some of the lists of people in corpora
.
On a technical level, for some reason I decided it would be a good idea to use cheapbotsdonequick to generate SVG files using the text for the different members of the household. This seems pretty silly in retrospect.
@badguystopper
https://twitter.com/badguystopper https://cheapbotsdonequick.com/source/badguystopper https://www.bbc.co.uk/news/av/world-us-canada-20817967
@mcdonnellbot
@mcdonnellbot
(stupid British politics joke;
inspiration),