Concept Taxonomy¶

The 326 concepts in the learning graph are grouped into 15 categories. Each concept belongs to exactly one category. The TaxonomyID is the 3-5 letter code used in learning-graph.csv. Categories are ordered from foundational material to the capstone projects.

Personal Agents and Architecture (FND)¶

TaxonomyID: FND
Concepts: 22

What a personal agent is, open versus cloud agents, the Open Study Buddy architecture, and the swappable-brain progression from a rule-based buddy server to open-source agent frameworks.

MicroPython Review (MPY)¶

TaxonomyID: MPY
Concepts: 20

The MicroPython skills the book assumes from the prerequisite course: REPL, modules, the machine library, interrupts, timers, memory limits and garbage collection.

Pico Kit Hardware (HWK)¶

TaxonomyID: HWK
Concepts: 22

The kit parts and their wiring: Pico 2 W, RP2350, round display, INMP441 microphone, MAX98357A amplifier, speaker, buttons and breadboard.

Round Display Graphics (DISP)¶

TaxonomyID: DISP
Concepts: 26

Driving the 360 x 360 round display over SPI: RGB565, frame size, why full-frame updates are too slow, direct region writes, clock faces, icons and the school mascot.

Buttons and Device Modes (BTN)¶

TaxonomyID: BTN
Concepts: 16

Mode, Up and Down buttons, debouncing, edge detection, menus, state machines, device modes and non-blocking event loops.

Digital Audio and I2S (AUD)¶

TaxonomyID: AUD
Concepts: 30

Sound, sampling, the Nyquist limit, the I2S bus, machine.I2S, audio buffers and frames, tone and stored-audio playback, and the self-hearing problem.

Signal Processing and FFT (DSP)¶

TaxonomyID: DSP
Concepts: 26

Time and frequency domains, the DFT and FFT, bins, windowing, magnitude spectra, band energies, loudness and spectrum displays.

High-Performance MicroPython (PERF)¶

TaxonomyID: PERF
Concepts: 20

Making MicroPython fast enough for real time: viper, inline assembly, the assembly FFT, typed arrays, allocation-free code, cycle counting and benchmarking.

Mic Calibration and Push-to-Talk (CAL)¶

TaxonomyID: CAL
Concepts: 14

Measuring the noise floor and speech level, signal-to-noise ratio, level gates, push-to-talk recording and utterance boundaries.

Wake-Word Detection (WAKE)¶

TaxonomyID: WAKE
Concepts: 24

Detecting 'Hey Pico': enrollment, feature extraction, template correlation, thresholds, false accepts and rejects, limits, and comparison with trained models.

Networking and Agent Link (NET)¶

TaxonomyID: NET
Concepts: 26

Wi-Fi, HTTP and JSON on the school LAN, chunked audio upload, speech services, WebSocket streaming, HTTPS, and handling network failure.

MQTT and Push Alerts (MQTT)¶

TaxonomyID: MQTT
Concepts: 22

Publish/subscribe messaging: brokers, topics, QoS, retained messages, last-will, umqtt, non-blocking check_msg, and connecting to Home Assistant.

School Agent and Calendar (SCH)¶

TaxonomyID: SCH
Concepts: 24

Connecting the agent to a school calendar via JSON then iCalendar feeds, test and due-date events, briefings, reminders and per-school customization.

Privacy and Responsible Listening (PRIV)¶

TaxonomyID: PRIV
Concepts: 18

Consent, parent and school approval, data minimization and retention, local versus cloud processing, threat models, mute and indicators.

Testing, Integration and Capstones (TEST)¶

TaxonomyID: TEST
Concepts: 16

Probe and self-test programs, wiring checks, logging, diagnosing silent microphones and blank displays, system integration and capstone projects.

Pedagogical Flow¶

The order above is the suggested teaching order. Personal-agent ideas, the MicroPython review and the hardware come first. Display, buttons and audio follow, then signal processing and performance. Calibration and wake-word detection build on audio and the FFT. Networking comes next, and MQTT builds on it. The school-agent, privacy and integration categories bring the pieces together.