DataScience webapp with Flask

Data Science webapp to show some of the capabilities of Flask and libraries such as sklearn, pandas, matplotlib, seaborn...

Capabilities:

Dataset upload

The webapp supports:

Dataset summary

This page will show:

Preprocessing

We can create a new dataset (it will be saved as CSV) with the following options:

Feature selection:

Null values and columns with a unique value:

Graphs

Available visualizations for the chosen variables:

Models

Models for Classification and Regression tasks. It does not support multiclass classification at this moment (extra code to manage some metrics and graphs)

Available Algorithms:

K-Fold Cross-Validation (3, 5, 10)

Standard Scaling (Yes, No)

Manual Feature Selection

Classification Tasks Output:

Regression Tasks Output:

Predictions

Model building (with the complete dataset) and prediction for a set of introduced values. The model will only include the variables with an introduced value. The available algorithms are the same that were mentioned in "Models". It also supports multiclass problems.


Some ideas for improvement:

VIDEO DEMONSTRATION

LINK TO YOUTUBE VIDEO