Figuring out what the heck I'm doing in 2025

So 2024 was… interesting. I didn’t start the year with any grand plans or a list of goals, but somehow I still managed to actually finish several projects that had been collecting digital dust for ages. While it feels good to have made progress, I’m thinking that maybe - just maybe - I might accomplish even more if I had some kind of plan? Revolutionary concept, I know.

Before I dive into my half-baked plans for 2025, let’s look at what I actually managed to do last year without adult supervision.

The evidence that I occasionally write code

If you look at my GitHub contribution graph for 2024, it’s pretty telling. I generated this using that GitHub Contributions Chart Generator thing, which is way more fun to play with than it should be.

The funny thing is how obvious it is when I started using GitHub for work stuff too. Suddenly my contributions went from “does this person know how to code?” to “maybe they should go outside occasionally.” It’s been pretty cool though I get to live on GitHub all day.

Night Sky Pi finally hit 2.0!

The biggest win last year was definitely getting Night Sky Pi to version 2.0.0. This project has been my baby for ages - it’s the system that automatically takes pictures of the night sky using Raspberry Pis and camera modules.

Version 1.0 took forever because I kept adding “just one more feature” (classic me). By the time I finally tagged it, it could:

• Take pictures all night long without me having to do anything
• Put all the images in folders that actually make sense
• Figure out by itself how long to expose each image based on how dark it is
• Save all the nerdy camera settings with each image
• Zip everything up so I can easily grab the data
• Send MQTT messages so I know what it’s doing
• Clean up old data so my Pi doesn’t explode
• Let me tweak settings without having to touch the code

For 2.0, I finally got around to making it properly installable. I ditched tox (which was a pain) and switched to Poetry for managing dependencies and building everything. The best part is I got it on PyPI, so now anyone can just pip install night-sky-pi and be done with it. No more “clone this repo and pray” nonsense.

The notification thingy that goes with it

I also built this little sidekick called NSP NTFY that takes all those MQTT messages from Night Sky Pi and turns them into phone notifications using ntfy.sh. Super handy because I don’t have to keep checking if things are working - my phone just pings me if the system starts/stops capturing or if something goes wrong.

It works, but it’s definitely the neglected sibling of Night Sky Pi. I need to:

• Make it use Poetry like the main project
• Get it on PyPI too
• Maybe make it more generic so it’s not just tied to Night Sky Pi

I’ve been thinking it could be useful for other MQTT stuff too, but we’ll see if I get around to that this year.

Harlow Bindicator - the least glamorous but most useful project ever

The Harlow Bindicator might sound boring - it just tells me when to put the bins out - but it’s probably the most practical thing I’ve built. I kept missing bin day and the other half was giving me grief about it, so I built this partly in self-defense.

It went through some painful iterations:

1. First version needed MQTT, Node-RED, and this sketchy SMS service. Way too complicated.
2. Second try was better - ditched most dependencies and used ntfy.sh directly. Added reminders for the night before (absolute game-changer).
3. Final version got the proper treatment with Poetry and PyPI publishing. It’s an actual command-line app now!

This one’s basically done unless the council changes their schedule format (please don’t). It’s been running for months without me touching it, which is the dream.

Smigle Lite - the theme powering this very blog

I forked the Smigle Hugo theme to make Smigle Lite because the original was close but not quite what I wanted. I basically:

• Ripped out the stuff I didn’t need
• Made it look a bit nicer (in my opinion anyway)
• Added some SEO stuff because why not

You’re looking at it right now! I should probably submit it to the Hugo Themes gallery, but that’s been on my “I’ll do it tomorrow” list for about 8 months now.

What I’m actually planning to do in 2025

Instead of making a massive list of projects that will mock me with their incompleteness next December, I’m sticking to just three main things. Quality over quantity and all that.

NSP Data Sender - The missing piece of the puzzle

Night Sky Pi captures all these great images, but getting them off the Pi is still a manual pain in the backside. I need something that will:

• Send images to my AWS S3 bucket automatically
• Also copy them to my NAS as a backup
• Maybe support FTP for good measure (though honestly who uses FTP anymore?)

The key thing is it should only delete stuff from the Pi once it KNOWS the data is safe somewhere else. The current time-based deletion makes me nervous.

If I can get this done, the whole Night Sky Pi system will finally be completely automated from capture to storage. Then I can think about doing something with all these images I’ve been collecting for years!

Cbujo - Because I can’t keep track of anything

Cbujo is my attempt at a command-line version of bullet journaling. I’ve tried using physical journals, fancy note apps, and todo systems, but nothing sticks. I live in the terminal anyway, so maybe this will actually work for me?

Right now it’s just a placeholder 0.0.1 release on PyPI (I squatted on the name, basically). I want to:

• Make a nice CLI with Click that doesn’t feel clunky
• Figure out a good way to store and manage the data
• Keep the commands simple and bullet-journal-like
• Add export so I can back it up or move between machines

Honestly, this is partly just an excuse to get better at building CLI apps in Python, which is surprisingly useful for work too.

Metmesh - The overly ambitious weather project

Metmesh is probably biting off more than I can chew, but I’m excited about it anyway. The idea is to collect weather data from all kinds of sources - APIs, web services, my own sensors - and do something useful with it.

For version 1.0, I want to:

• Build a system that can pull from various weather data sources
• Normalize all that data into a standard format
• Store it properly in a database
• Calculate hourly/daily/monthly stats
• Group data by location in a way that makes sense
• Create some kind of API so I can actually use the data

I’m starting with basic weather stuff - temperature, humidity, pressure, etc. - but the architecture should support adding more exotic measurements later.

Fixing my sad excuse for a workspace

Let’s be honest - my current workspace is depressing. It functions, but it’s about as inspiring as a dentist’s waiting room.

Just look at that. Issues include:

• Blinding white surfaces everywhere (I feel like I’m coding in a hospital)
• Unfinished brick and block walls (industrial chic or just lazy? you decide)
• Bare concrete floor that’s impossible to keep clean
• Zero personality or visual interest

In 2025, I’m going to make this place less soul-crushing:

1. Cover up some of this brick with wood panels or something
2. Add some actual color that isn’t “sterile white” or “construction site”
3. Put down some kind of flooring that doesn’t make my chair wheels get stuck
4. Replace those horrible fluorescent lights with something that doesn’t make me look like a zombie on video calls

I spend too much time in here to keep pretending the environment doesn’t matter.

Getting my stuff organized (for real this time)

Half my project time gets wasted looking for that specific screwdriver or USB cable I KNOW I have somewhere. My “organization system” is basically various piles that occasionally get shuffled.

I’m going to:

• Finally figure out what tools and parts I actually have
• Come up with some kind of categorization that makes sense
• Stop keeping things I’ll “definitely use someday”
• Set up my desk so I can actually start working without 15 minutes of preparation

I want to get to a point where I can think “I should work on X” and actually be doing it five minutes later, not searching through boxes and drawers for an hour first.

Making my Pi cluster less annoying

My current Raspberry Pi setup is mounted on those IKEA Skadis pegboards, which seemed clever at the time. The problem is I’m using PoE hats to power them, and those fans are LOUD. Like “can’t-hear-myself-think” loud. Also impossible to have on during Zoom calls.

New plan:

• Get a small rack to mount everything properly
• Ditch the PoE hats and just use regular power supplies
• Sort out the cable mess once and for all
• Figure out a cooling solution that doesn’t sound like a jet engine
• Make the whole thing look less like a science experiment gone wrong

The PoE setup was an expensive mistake - those hats cost me about £30 per Pi, plus I needed special PoE switches. Regular power is cheaper and quieter, which seems like a win-win.

Writing stuff that people might actually want to read

I’ve been sporadic with posting here, but when I do write something, it tends to be either super detailed project logs or quick updates that don’t say much. For 2025, I’m thinking quality over quantity:

1. Proper documentation for my projects that might actually help someone else
2. Tutorials for stuff I figured out that wasn’t well explained elsewhere
3. Posts about technical problems I solved (future me will thank me when I inevitably hit the same issue again)

I’m not going to promise “one post per week” or anything like that, but I do want to:

• Write at least one substantial post each quarter
• Document each major project properly
• Write up any interesting technical solutions while they’re still fresh in my mind

So that’s the plan… we’ll see how it goes

By the end of 2025, I’ll consider it a success if I’ve actually completed the three main projects rather than having ten half-started ones. And if my office doesn’t look like a storage unit with a desk in it, that would be a nice bonus.

These projects all build on stuff I’ve already started, which hopefully increases my chances of actually finishing them. They’re practical things I’ll actually use, not just shiny new tech for the sake of it.

Most importantly, I’m trying to be realistic about time. I’ve got a full-time job, other commitments, and an attention span that… hey, what’s that over there? Anyway, by focusing on just a few things and making my workspace less chaotic, maybe I’ll actually get somewhere this year.

Let’s see how this post ages, shall we?