Data Modeling GUI

Application for small data modeling tasks with GUI

Features

With this application the user is able to load data, preprocess and explore it (to some degree), and make predictions for regression or classification tasks.

• Select Feature types (inputs/outputs, numeric/categorical)
• automatic handeling of missing data
• automatic data summary and visualization
• classification and regression tasks (with models below)
• custom model parameters
• data and model export (.csv, .pkl)

Models

Classifiers

• Decision Tree Classifier
• Random Forest Classifier
• Gradient Boosting Classifier

Regressors

• Linear Model
• $k$-Nearest Neighbors Regression
• Decision Tree Regression
• Gaussian Process Regression
• Neural Network
• Support Vector Regression
• XGB Regressor

Possible Future Features

Better support for categorical data: use scikit ColumnTransfromer! Include more models, add a real progressbars, and add cross validation.

Dependencies

The following packes are required

• tkinter
• sklearn
• xgboost
• numpy
• pandas
• pickle
• ctypes
• matplotlib
• seaborn
• pandastable
• inspect
• webbrowser

Feedback is appreciated!