Machine Learning Workflow App

Project Overview:

An interactive web application built with Streamlit that provides a complete end-to-end Machine Learning workflow — from data upload and preprocessing to model training and evaluation.

️ App Features:

Data Upload & Exploration:

Upload CSV files, view data summaries, and basic statistics.

Preprocessing:

Handle missing values (Simple/KNN/Iterative Imputer), encode categorical variables, detect and treat outliers (IQR, Z-Score, Winsorization), scale features (Standard, MinMax, Log, Power, Polynomial), and perform feature selection (PCA & RFE).

Visualization:

Generate multiple plots (Pie, Bar, Histogram, Heatmap, Pair Plot, Box, Violin, Scatter, KDE, etc.) to explore data insights.

Model Training:

Supports Supervised (Classification & Regression) and Unsupervised (Clustering) learning.

Models include: Logistic Regression, Decision Tree, Random Forest, SVM, KNN, Linear Regression, SVR, and KMeans.

Model Evaluation:

Dynamic metrics based on task type:

Classification: Accuracy, Precision, Recall, F1-score, ROC Curve, Confusion Matrix, Classification Report

Regression: MSE, MAE, R²

Clustering: Silhouette Score & cluster visualization

? Tech Stack:

Python – Streamlit – Pandas – NumPy – Scikit-learn – Seaborn – Matplotlib