Last week

Weeknotes

After six months of trying to write as-I-go-per-topic weeknotes, rather than one-a-week-all-topics weeknotes, I declare the experiment a failure, and I'm going to switch back to per-week. The temptation to wait just a little longer on each per-topic note to get one more thing done was too much, and so the cadence of posting was too low for weeknotes, and I was effectively doing a poor form of long-form blogging. So I'm going to switch back to regular short-form weeknotes with all topics in them, and then occasionally try to do wrap-up long form posts. I made the change because I'd been aware that the weeknotes were kind of a pain to read with so much different stuff in them, but ultimately weeknotes are really a tool for me, not for others. I mean, I like it if others find them useful, but they are for me a way to keep track of what I did, and I've lost that with the change, so back I go.

EMFCamp

I spent the better part of four days in a field at EMFCamp, which has been fun. Not a huge amount technical to report from that, just it was good to chat and reconnect (and freshly connect) with people. Like a few others, I skipped the fixed schedule, and just spent them time wondering around the site and having conversations with people instead.

Habitat mapping with Tessera

After talking about it for a long time, I've finally spent some time trying to do some actual practical work using Tessera to build up some habitat maps, in collaboration with Ali and Aneesh. I actually have in the wings some long form posts on Tessera, so I won't go into details here - these were some "weeknotes" that ballooned into detailed epics, which I wrote when I did a bit of using Tessera to explore the Swedish land cover maps earlier in the year. I can hopefully dust them off as this work progresses to explain what it is I'm trying.

The main focus last week was just getting up to speed on some of the specific challenges of using Tessera to build maps using the messy data that I'll be using this time, and reading papers on what others have done before in this space. I have written a very naive end-to-end pipeline that isn't going to make great results, but should help me understand what it is I actually need, and where there are going to be data gaps that need filling.

The one thing that did catch me out is how big the Tessera data is. I work remote and have a powerful Mac Studio on which I do most of my analysis work, and I've kinda got used to being able to work that way. I'd planned to pull data for the UK and Madagascar for two years, but just the UK for one year was threatening to take 2 days to download here, and so I've switched back to using the lab machines for now for this work. Once I found my feet I suspect I'll just take a large disk to Cambridge with me and copy the bits of Tessera I need and bring them back with me :)

Tangled CI

Despite being very happy with Tangled I found in practice I unfortunately still end up having to run a Github-first life due to my reliance on Github Actions for quality control: even for projects where it's just me I make PRs against myself so that unit tests and ran and checked, and Tangled doesn't have that out the box. To that end I've dragged my old PC out of retirement and put linux on it (it was running Haiku :) so I can try set up a spindle, Spindles being the CI part of the Tangled ecosystem. If that works, then I wonder how hard it would be to have a spindle on my Mac Studio run my pipelines for me if I nominate some blessed branch.

This week

• Get my little Tessera pipeline running end-to-end
• David Coomes' group are having a half-day meeting to review various Tessera/Habitat projects that are in progress, so I plan to attend that to try and get up to speed on the bits I don't know, and get feedback on the plans myself, Ali and Aneesh have pulled together.
• Find some book to listen to on the drive to/from Cambridge. On the previous Cambridge run I listened to We Are Not Machines by Sarah O'Conner. It was sufficiently interesting I bought the paper version so I can look up some of the references in it and blog about them here.