README.md

Introduction to Sensors

💦🌱 lecture slides ☀🌳

Time series modeling exercise

Example models and data

Use greenhouse data to create a Python model based on one of four available example machine learning models (Kalman, SARIMAX, RNN, or SVR), or create a new model from scratch.

Use at least two sensors to predict the output of one other sensor.

Micropython Lunar Lander filter exercise

Micropython filtering

One of the main differences of coding for an MCU is that they offer very few resources. It is generally not possible to use large Python libraries such Numpy, Scipy, SkLearn or PyTorch.

The Lunar Lander assignment requires you to implement a filter while working around these limits.

MQTT streaming data analysis exercise

MQTT exercise

Run the minimal MQTT chat example, and chat with your follow students. Work together with your team: let one of your team publish data and another(s) receive data, generated by and analyzed with https://github.com/jmaces/statstream

Scailable edge computing exercise

Scailable 101

Fit a Python Sklearn SVR model on the time series data from the first exercise, convert it to WebAssembly, and make it available as a rest endpoint using the the Scailable edge computing platform. Follow Scailable 101 to get started.

TTGO HiGrow exercise

TTGO HiGrow with MQTT Cayenne integration

An exercise with the TTGO HiGrow: a $10 microcontroller based device with several plant related sensors. Use your HiGrow Device to measure the environment of one of your own house plants for a week, then analyse the resulting data with a time series model.