TrEnki application

TrEnki is a web application aimed at supporting the prediction of faults in pipes of drinking water distribution networks using machine learning techniques. The application is deployed on the university's servers and allows users to create, manage, and analyze work environments and studies from data sets with a specific format.

Through a simple and efficient interface, the user can execute the main stages of a data science project: cleaning and filtering the data, analyzing its distributions, preparing the training and test sets, training new models or loading previously trained models, validating performance and generating predictions. The results can be visualized through an interactive map, making easier the spatial interpretation of the predictions and supporting the prioritization of actions.

TrEnki is a flexible tool that allows working with different models and hyperparameter configurations, carry out optimization processes and export intermediate results for its external analysis or integration with other distribution networks.

It is shown below, in detail, with images showing the views corresponding to the main workflow:

The application can be found on the next website: https://esaiict.upc.edu/trenki/ 

In order to be able to use the application, it is required to fill the form, accesible by pressing the button below:

Form