All activities
24 hands-on activities, ordered by chapter. Each is meant as a ready-made double lesson — learning goal, materials, step-by-step, worksheet with solutions and an extension. About half run entirely in the browser.
M Master extension (advanced) R Reflection D Discussion
Make heart-rate variability visible — how tension changes the rhythm of your pulse.
Measure your own class chat: response times, balance of contributions, giving and taking.
Analyse a WhatsApp chat for the five roles with the ready-made web edition.
Have a language model write the analyzer in four steps (vibe coding).
From word-counting to a language model: fine-tune a BERT and catch the Leech with Sentence-BERT.
Learn how to export your WhatsApp conversations, upload them to Beecome, and understand your relationship analysis.
Network analysis of a real open faculty: find the queen bee by betweenness (networkx).
Recognise emotion from the face in real time (webcam, HSEmotion) — and feel the unease of it.
A memoryless model applied many times: record and plot your own emotional course.
Retrace "Your Face Mirrors Your Deepest Beliefs" — deliberately as reflection, ethically delicate.
Measure pitch, melody and tempo — what the voice reveals that the words don't (librosa, Whisper).
Build the sensor and electrode, see the first voltage curve on screen.
Observe how a plant responds to stimuli — with a clean control.
Record the plant's response to tones — and separate vibration from sound.
Does it respond to my presence and my mood? The big open question — posed cleanly.
The two-stage sound AI: the kind first, then the mood in the meow (real data).
Let a model guess an animal's mood from a photo — and check it yourself (pure inference).
Make a 24-hour recording and find circadian rhythms (FFT).
Spectrogram → a pre-trained network reads the mood: pure inference, no training of your own.
Draw the crowd's emotion curve from eight years of headlines (VADER).
Lay the mood against the Dow Jones — does it predict it? (the Mule).
Compute the symbiont shares per relationship and see that the twin depends on the person opposite.
Fine-tune a mini language model on your own texts — your echo, not your I.