Gwmngilfen @gwmngilfen@fosstodon.org
#FOSS evangelist, #linux, #rstats, #datascience, #baking, #walking, #gaming, and other randomness.
Work: Community data scientist for https://twitter.com/ansible and https://twitter.com/redhatopen
Work stuff happens mostly on https://twitter.com/gwmngilfen, personal stuff here (because Fedifolks are the best :p)
Joined Jul 2018
Pinned toot
Alright, time for a new home. Fosstodon has been recommend to me by quite a few people as the place that matches my interests, so, uh, here I am!
For those not following me over from mastodon.org.uk, I'm a FOSS community manager, wannabe data scientist, sysadmin and devops enthusiast. Currently very interested in #rstats and #data #analysis, but will happily talk endlessly about pretty much anything.
Gwmngilfen
boosted
Hey #devops peeps: time to upgrade All The Things. https://www.cisecurity.org/advisory/a-vulnerability-in-gnu-c-library-could-allow-for-remote-code-execution_2020-105/
Finally starting to think about consolidated self-hosting platforms. Would like to do some evaluation, but I'm a bit out of the loop - I'm aware of @yunohost and Cloudron.io (which is paid), what other options are there?
(Honourable mention for FreedomBox too, but I *think* that's not massively extensible in the way something like Yunohost is, right?)
Fire away, Mastodoons!
An evening spent building infrastructure for my local community. Time well spent, feels good. Should do more of this.
Gah. Cracked the screen on my OnePlus with a careless drop. Not serious, hairline fracture over the back button, but still, it gives this a definite expiration date.
@cybette still enjoying the F(x)Tec? Would you recommend?
Been a while since I bought a book, time to play "covid-catchup"...
My secret stats book dealer recommended this to me by Richard McElreath, and two chapters in, I'm loving the style! Thought provoking without being preachy, detailed. V.good for self-taught folks like myself!
Beaches - as they should be. Even without Covid19, I prefer an empty beach :)
entirely self-serving self-promotion
Was super happy to be a guest on the Scotland Open Source User Metups podcast this week! Lots of chat about communities, explotation of dev work, education, charities etc. If that's your bag, have a listen?
https://scot-osum.gitlab.io/podcasts/2020/07/07/sosum-podcast-s01-ep03.html
Big thanks to @samusaran for inviting me and doing all the editing :)
Gwmngilfen
boosted
Gwmngilfen
boosted
infosec, privacy, tiktok
TikTok has been reverse engineered, and found to be surveilling its users extensively, collecting data including clipboard contents, phone hardware, other installed apps, GPS data, and whether or not the phone is jailbroken.
All of this is hidden behind extensive obfuscation, and is remotely configurable.
Gwmngilfen
boosted
Google Analytics is the most prevalent tracker in the world. During our last crawl, we found it on 64.2% of web sites:
https://better.fyi/trackers/google-analytics.com/
Developers: every time you add Google Analytics to a site, you are helping build the global panopticon. Stop it!
Gwmngilfen
boosted
My post today reflects some thoughts I've been thinking about as I look for better ways to approach privacy myself and look at some of the really bad takes I've seen in the area. https://chriswiegman.com/2020/06/privacy-empathy-and-the-zero-sum-game/
Gwmngilfen
boosted
> 1500 #games at a nominal cost (which you can redefine) for an amazing cause - donation towards NAACP Legal Defense & Educational Fund and Community Bail Fund. https://itch.io/b/520/bundle-for-racial-justice-and-equality
Thanks to #itchio and all the creators!
h/t @gwmngilfen
I wonder if Victorian metallurgists felt the way many of us data folk do today?
Surely a few people must have been saying "look, I know lead is super easy to work with, and makes plumbing a doddle, but, uhh, its poisonous. This isn't going to end well..."
I will be mostly spending today bring outraged at the things I'm *required* to sign up to for my children's childcare/education, with no chance of reviewing the Terms & Conditions / Privacy Policies first.
Sad to see local councils outsourcing to Google, especially Google Classroom. I wish we could see more #foss consideration in tendering processes - keeping control of the data you have on residents of the area has to have a value? Self hosting doesn't have to be hard.
Le sigh
@rstats Dplyr 1.0.0 is out!
https://www.tidyverse.org/blog/2020/06/dplyr-1-0-0/
Ludicrous amount of stuff, bu my pick has to be summarise() being able to handle creating new rows as well as new columns - which can replace all sorts of nasty patterns involving do() or map() ...
(via https://twitter.com/hadleywickham/status/1267546628224884736)
Current #COVID19 avoidance method - being completely obsessed with Oxygen Not Included (https://store.steampowered.com/app/457140/Oxygen_Not_Included/)
I love a good mix of planning, exploration, & deep systems, and under the cartoony surface, this has it all. Eg. I'm currently building transmission & distribution power lines because basic wiring only carries 1kW...
I bounced off it a while back, but I went back to it this week fsr, and now it's clicked. I just got to the Oil Biome - onwards!
@rstats your help please people!
I have a frame of ids and start/end dates, eg
acc,sys,start,end
1,1,2020-04-01,2020-04-21
...
I need to group by `acc`, and then calculate the total active systems over the weeks the account was active at all, eg
week, acc, total_sys
2020-04-11,1,17
...
I have a solution that works, but takes ~1hr to run (initial df is 1.2mil rows). Is that ~right, or should I be able to speed it up?
Can share the code, ofc
/me goes to investigate the `furrr` package...
#FOSS evangelist, #linux, #rstats, #datascience, #baking, #walking, #gaming, and other randomness.
Work: Community data scientist for https://twitter.com/ansible and https://twitter.com/redhatopen
Work stuff happens mostly on https://twitter.com/gwmngilfen, personal stuff here (because Fedifolks are the best :p)
Joined Jul 2018
