Just surpassed 1,300km of running for this year. I am very pleased with this, and it didn't really feel like effort (it's just part of my day now)
I'd love to shoot for 1,400 before year end. That'd mean some bigger (~20km+) runs this week, between all the holiday stuff we got going on.
Won't push myself, but that'd be rad. Will probably blog about the year regardless.
We recently published our Air Pollution Data Explorer.
Explore historical and recent emissions of air pollutants across the world: https://ourworldindata.org/launch-air-pollution-explorer
Idea: Fine tune stable diffusion to understand how to generate imagery that can introduce traffic calming infrastructure to a photo of a street.
"A picture is worth a thousand words". Show neighbours what's possible, talk to councilor, and keep pushing for change?
Possibly helpful .. but also an excuse to try something interesting with evolving tech :)
These are old, but I found GitHub issues from the sidekiq author on how to further reduce. Ultimately with running a single-user instance, I'm not going to need such a level of complexity/scale as a typical server would
Will look at this next.
Setup my own Mastodon instance because if I'm going to go federated, that feels appropriate?
Running it on a tiny 1 vCPU / 1GB Digital Ocean droplet which I boosted to 2GB during compilation from source (some OOM otherwise)
Sidekiq (https://sidekiq.org/) which is the included job queue, was also running with many workers to begin. I sliced that down and was able to resize my droplet
I imagine there's a lot of potential for making the single-user instance deploy easier!
Transit projects in North America are typically more expensive and lengthier in time.
Why? Seems to be: Partly labour costs, but mostly *a lot* of dawdling and expensive consultation for implementing technology that is pretty standardized