Skip to content

Tutorials

A series of seven notebooks in notebooks/, meant to be read in order. Each one runs on its own and builds its own data, so no plant access is needed.

Notebook Question it answers
00_quickstart The whole workflow in one pass
01_explore_measurements What is in the data, and where is it broken?
02_model_vs_measurement How far off is an uncalibrated model?
03_first_calibration Fit one parameter until the curves meet
04_which_parameters_matter Which parameters is it worth fitting at all?
05_train_test_and_residuals Does the fit hold on data it has not seen?
06_online_recalibration Following a plant that drifts

Notebook 00 runs in Google Colab without a local installation:

Open In Colab

The others expect a local install. They share notebooks/demo_plant.py, which builds the small plant and generates the measurement records.