Decision compiler
OfOne
An open-source decision compiler: a hard question becomes a typed, validated map, and the prose you read is a rendering of that map.
Work/Signals, not vibes
EEG meditation toolkit
The EEG meditation toolkit is Python tooling that turns a raw consumer-headset recording into a readable account of a meditation session. Filtering removes artifacts, power is measured in each brainwave band, and a depth score from 0 to 100 rolls up with one of four labels: alert, light, moderate or deep. A synthetic data generator ships with it, so the whole pipeline runs without a headset and without handing over a personal recording.
Open sourceExperiments
The problem
People who meditate have almost no way to tell whether a session went deep or shallow, other than how it felt. Consumer EEG headsets record the signal, but the raw signal is noise to a human eye. The gap is not the sensor. It is the analysis in between.
What I built
A Python toolkit that takes an EEG recording and reports what happened during a meditation session. It filters the signal, removes artifacts, measures power in each brainwave band, and scores meditation depth from 0 to 100. It can also walk a whole session in sliding windows, so you can see where attention drifted and where it settled.
It ships a synthetic data generator, so anyone can run the whole pipeline without owning a headset or handing over their own recordings.
How it works in plain words
Brain activity gets sorted into frequency bands. Delta is the slowest, from 0.5 to 4 cycles per second. Theta runs from 4 to 8. Alpha, Beta, and Gamma go up from there. Deep meditative states show more of the slow bands and less of the fast ones.
The toolkit measures how much power sits in each band, then computes ratios between them. A higher Delta to Theta ratio points to deeper states. The Theta to Alpha ratio tracks the move from alert to meditative. It also reports spectral entropy, which is a measure of how disordered the signal is, and the dominant frequency.
Those numbers roll into one score and one of four labels: alert, light, moderate, or deep. It draws the whole thing as a multi-panel chart with a spectrogram, the power spectrum, and the band distribution.
Honest limits
This is a signal-processing toolkit, not a clinical instrument and not a validated study. The four state labels are thresholds taken from published band conventions, and no control group was used to check them. No accuracy claim is made here.
Status: dormant. The last commit was 2025-11-11, about nine months before this page was drafted.
Pending from Jameslabel this "archived" or "dormant, open to revival"? Also: was this run on your own recordings from a specific headset, or on synthetic data only? If real recordings exist, name the device.
Stack
Python 3.8 or newer. NumPy and SciPy for filtering and power spectral density. Matplotlib for the multi-panel plots.
Stack
Python 3.8+NumPySciPyMatplotlib
Related work
Spend telemetry
plimsoll
A local-first collector that joins AI coding spend to merged pull requests, so cost per shipped outcome stops being a guess.
Findability, measured
The visibility platform
A multi-tenant platform that measures how findable a business is in search and in AI answers, and prints the evidence under every score.