Interactive ML & Data Preprocessing Web App

Developed a modular and interactive web application using Streamlit for end-to-end machine learning workflow,

including data profiling, preprocessing, visualization, and model evaluation.

Key Features:

• Uploaded and profiled datasets using ydata-profiling

• Handled missing values, encoding, outliers, normalization, and duplicates

• Visualized data with pie charts, bar charts, strip plots, and correlation heatmaps

• Trained and evaluated models (Decision Tree, SVM, KNN, Naive Bayes, Random Forest, KMeans)

• Applied both train-test split and K-Fold cross-validation with real-time metrics and confusion matrix

visualization

• Enabled user control over feature selection, target column, encoding strategy, and scaling