Pandas Supersónicos

Inspiration

The project is inspired by the well-known company Schneider Electric.

What it does

We created a binary classifier model that predicts the results of an opportunity (a sale), classifying them as losses and wins according to the attributes of each opportunity. After that, we have analyzed these predictions in order to understand what guided the model to classify on them in this way.

How we built it

We used Sklearn, a Python library, to train the model using XGBoost. And then we used explainability techniques such as SHAP, ICE, PDP and ALE to interpret the influence of each variable.

Challenges we ran into

We ran into Schneider Electric Challenge because we were curious about the explainability techniques, which we have never heard of.

Accomplishments that we're proud of

We understand some of these techniques that will be extremely useful in future projects.

What we learned

We understand some of these techniques that will be extremely useful in future projects interpreting data from a dataset. Also we have learnt about the model XGBoost.

What's next for Pandas Supersónicos

We can investigate more in deep with the variables from the dataset, since we spend most of our time in understanding how does the explainability techniques works.

Built With

• castellano
• github
• python
• r
• sklearn