This project provides code to analyse data from the Chess Watch project of the river Chess. Data from Chesswatch are available at https://rhysh.shinyapps.io/ChessWatch Here, we provide extracts of that data, together with the code to reproduce most of our findings: The data should be put in the folder "data", figures will be exported to the "pics" folder. Pre-processing is included in notebook 0, using constant values in place of the Chesham data (see below for details). Each remaining notebook focuses on one aspect of the paper (as they appear in the paper): 1. Exploratory data analysis (trajectories, PDFs, moments, ...) 2. GAM 3. Boosted Trees and SHAP There are two versions of notebook 3: One where the model and its hyperpameters still have to be determined and one where the parameters are fixed to the ones shown in the figures of the paper. When applying this code to different data, determine new hyperparameters, use new hyperparameters. IMPORTANT NOTE: The data from Thames Water from the Chesham plant cannot be shared here. To still make the code executable, we provide mock-data, where the outflow value is always 100 instead of the real values. Please contact Thames Water to get access to the Chesham data yourself. All analysis not using the WWTW outflow explicitly (most of the analysis) can be fully reproduced with this code and data. When using this, please also cite the article "Boosted tree with SHAP analysis reveals the influence of treated effluent on electrical conductivity and temperature dynamics in an urbanised groundwater-fed river." by Benjamin Schäfer, Christian Beck, Hefin Rhys and Catherine M. Heppell.