Recently, I copied and pasted a day’s worth of raw coding activity data from WakaTime (if you’re not familiar with WakaTime, check out this post). To my surprise, the result I got back was extremely useful to guide my work. Here are some examples:
Practical advice (based on what you actually did)
Astro – cut duplication now
You opened/edited multiple pages (sites/*.astro, themes/*.astro, plugins.astro). Centralize the data and generate the pages through mapping:
- Create
/src/data/themes.jsonand/src/data/plugins.json. - Make
themes.astroandplugins.astrorender cards from these JSONs. - Have a single component (
ThemeCard.astro/PluginCard.astro) + aBaseLayout.astro.
Focus and measurement (with WakaTime data)
Today’s signal shows lots of reading/inspection and little effective writing (~20–25% of heartbeats appear as is_write: true).
- Try 45–60 min blocks with a single goal per block (e.g.: “themes.astro reads from themes.json and renders 3 cards”).
- Create a Goal in WakaTime for “Astro write time” and aim for >40%
is_writein those blocks.
Next steps (hands-on tasks)
themes.json+ThemeCard.astro+themes.astromapping data.- Adjust
deploy.ymlwithastro check + smoke (curl). - Create
bin/new-sail.phpto generatesail/YYYYMMDD.phpfrom a stub. - Standardize the README of Belo Bigode with fixed sections (reusable in other repos).
It even generated sample code and pseudo-code.
Since I really enjoyed this experiment, I decided to automate it. But right at the start of development, I ended up creating a relatively simple service, where the user simply connects to WakaTime, and NextCommit does the rest.
After a few hours of coding (around 5 logged by WakaTime), I released early access to a few people so they could provide feedback on the functionality itself.
For now, since it’s still an MVP, I need to run some scripts manually and risk losing a few queries due to timeouts. But I’m already working on improvements for the service.
If you’d like to check it out, visit: https://nextcommit.andreyrocha.com/