Welcome to the SuperDataScience Community Project!

Welcome to the Social Sphere: Student Social-Media Behavior & Relationship Analytics repository! 🎉

This project is a collaborative initiative brought to you by SuperDataScience, a thriving community dedicated to advancing the fields of data science, machine learning, and AI. We are excited to have you join us in this journey of learning, experimentation, and growth.

To contribute to this project, please follow the guidelines avilable in our CONTRIBUTING.md file.

Project Scope of Works:

Project Overview

Social Sphere explores how social-media habits relate to students’ academic performance, sleep, mental health, and relationship dynamics. Using an anonymized cross-section of 16- to 25-year-olds from multiple countries, the project will:

• surface global insights on usage intensity, platform preference, and well-being;
• predict relationship conflicts and self-reported addiction levels;
• segment students into behavior-based clusters;
• deliver an interactive Streamlit app for researchers, educators, and mental-health advocates.

Link to Dataset: https://www.kaggle.com/datasets/adilshamim8/social-media-addiction-vs-relationships

Project Objectives

Exploratory Data Analysis

• Profile demographics (age, gender, academic level, country) and social-media metrics.
• Visualize links between Avg_Daily_Usage_Hours, Mental_Health_Score, and Sleep_Hours_Per_Night.
• Assess country-level and platform-level differences.
• Handle missing values, encode categoricals, and flag potential biases.

Model Development

1. Predictive Tasks

   • Conflict Classifier: predict Conflicts_Over_Social_Media (high vs. low) from usage patterns, addiction score, and platform.
   • Addiction-Level Regressor: estimate Addicted_Score as a continuous outcome.

2. Clustering

   • Group students into interpretable segments (e.g., “high-usage high-stress”) with K-Means / HDBSCAN on scaled features.

3. Evaluation

   • Classification: Accuracy, Precision, Recall, F1, ROC-AUC.
   • Regression: MAE, RMSE, R².
   • Clustering: Silhouette Score, qualitative segment validation.

4. Hyperparameter Tuning

   • Grid / random search to optimize tree-based and gradient-boosted models.

Model Deployment

• Streamlit Web App

  • Dashboards (usage heatmaps, correlations, country comparisons).
  • Live predictor for conflict likelihood and addiction score.
  • Cluster explorer with segment descriptions.

• Host on Streamlit Community Cloud (or similar) with public README and user guide.

Technical Requirements

Tools and Libraries

• Data Handling: pandas, numpy
• Visualization: matplotlib, seaborn, plotly
• Machine Learning: scikit-learn, xgboost, lightgbm, mlflow
• Clustering & Dimensionality Reduction: KMeans, HDBSCAN, PCA/UMAP
• Deployment: Streamlit

Environment

• Python 3.9+
• Virtual environment (conda or venv)

Workflow

1. Setup & EDA (Week 1)

   • Set up GitHub repository, project structure, and virtual environment.
   • Data profiling, outlier checks, visualization, and bias assessment.

2. Feature Engineering & Model Development (Weeks 2 - 4)

   • Encode categoricals, scale and transform numerical features, derive interaction features.
   • Train and tune classifiers and regressors.
   • Build clustering pipeline.
   • Experiment tracking using ML Flow

3. Deployment (Week 5)

   • Build Streamlit UI (Insights, Predictors, Clusters) and deploy to the cloud.

Timeline

Phase	Core Tasks	Duration
1 · Setup & EDA	GitHub repo, project structure, virtual environment, data profiling, outlier checks, visualization	Week 1
2 · Feature Engineering & Model Development	Design, train, experiment with models.	Weeks 2 - 4
3 · Deployment	Build Streamlit UI (Insights, Predictors, Clusters); deploy to cloud	Week 5