This program is developed by Emil Sejer Pedersen, Frederikke Rønne Westergaard, Andreas Work, Lucas Joachim Kolding and Marie Margrethe Riskær Povlsen.

The program has been developed as part of our bachelor project on machine learning explainability.

To run the code it is necessary to import the following libraries:

Numpy
Pandas
Plotly
Dash
JSON
MatPlotLib
Seaborn
Scikit-learn
Graphviz
SciPy
OS
RANDOM
LIME
PIL

Furthermore to generate the graphviz elements, graphviz should be installed on the computer.

Whatever dataset the user wants to use can be inserted in the Datasets folder, and from there the dataset can be cleaned using the methods from the DataClean class.

In the dashboard the ML-classes are called, to set the settings for these please navigate to the DashboardController class.

When the code is run, it can take a while before the dashboard is live. This is due to the pictures the program generates. When using a dataset of about 2000 lines, it takes us 1-2 minutes to launch the dashboard.


The program is meant to be used with the above mentioned programs and imports. In addition, we will point out that the explainable models can be misinterpretted, and renounce all responsibility of potential misuse.