This codebase gathers the different notebooks used to exploit the data from Covidom to uncover patterns in PCR results
Install packages from requirements.txt
.
-
univariate_analysis.ipynb
perform univariate analysis of testing factors and PCR results. It produces the plot from Figure 1 and 2 -
train_tree.ipynb
produces Figure 3 and related decision tree figures, as well as temporal figures. -
eval_tree.ipynb
produces Figure 4.
The data from data/shuffled.csv
are generated data with the same columns as those that we use for analysis -- it is therefore fake data, but can be used with our scripts.
The real data is available upon request. It has to be analysed on APHP servers.