A package for creating and handling river hygiene models

The kwb.flusshygiene package provides functions in three major categories: model handling, model creation and model prediction.

Model handling

river_model_prediction: is the main function in this package. It uses all of the following functions from within.
get_paths: reads a serverpath library for easy directory accessing
search_existing_models: searches saved R-objects in the river directories.

build_new_model: again a overhead function for model creation. Asks the user whether or not the new model shall be saved as R-object.
import_riverdata: reads all river data from a directory.
build_model: is a small function handling model creation and invokes stan_lm for model building.
ask_for_variables: asks the user which variables shall be included in the model. Creates plots as a side effect.
process_model_riverdata: processes a data.frame with the necessary data for the data argument in stan_lm
create_formula: creates a hygiene formula out of the variables with the form e.coli ~ Q * (R + Ka)

predict_quality: is the overhead function for the prediction. It also invokes posterior_predict.stanreg
get_newdata: gathers the latest data for prediction.
print_latest: prints the prediction of the latest day.
plot_predicted_quality: plots a whole season with quality assessment.

unroll_physical_data: unrolls a list with data with lagday combinations to 5 days (default)
correlation_scatterplot: plots a scatterplot matrix of the unrolled physical data together with correlation values.

plot_stan_model: plots a stan_lm posterior predction with quality assessment.
plot_data_overview: plot data overview
plot_hygiene_overview: a statistical hygiene data overview
plot_q_overview: plot a overview of all q values
plot_rain_overview: plot a monthly overview of all rain gauges