Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 

Repository files navigation

EDA Dashboard — 9 Real-World Datasets

An interactive Streamlit dashboard for exploratory data analysis across 9 diverse datasets covering banking, finance, fraud detection, insurance, transportation, retail, and gaming.

Live demo: share.streamlit.io (deploy from this repo)

Datasets

# Dataset Size Target
1 Bank Marketing 41,188 rows Term deposit subscription prediction
2 Credit Card Fraud 284,807 rows Fraudulent transaction detection
3 Loan Default Risk 10,000 rows Loan default prediction
4 Medical Cost Insurance 1,338 rows Premium cost drivers
5 NYC Taxi Trips 100,000 rows Trip fare & tipping analysis
6 Online Retail Sales 541,909 rows Revenue & customer analytics
7 S&P 500 Bank Stocks 11,286 rows Stock price & correlation analysis
8 Steam Games 40,833 rows Game pricing & ratings
9 Video Game Sales 16,598 rows Global sales by platform & genre

Features

  • Data Quality — missing values, cardinality, duplicate detection, dtype overview
  • Statistical Summary — descriptive stats, distribution histograms, correlation matrices
  • Outlier Analysis — IQR-based detection, box plots, outlier percentages
  • Class Balance — target variable distribution (binary classification datasets)
  • Feature Engineering — derived features per dataset (age groups, loan-to-income, tip percentage, price tiers, etc.)
  • Key Insights — data-driven findings for each domain

Tech Stack

  • Python 3.14 — Pandas, NumPy, Plotly
  • Streamlit 1.58 — interactive multi-page dashboard
  • EDA Utilities — shared module for data quality reports, statistics, outlier detection, and correlation analysis

Local Setup

pip install -r dashboard/requirements.txt
streamlit run dashboard/Dashboard_Home.py

Deployment

This repo is ready for Streamlit Community Cloud. Point the main file to dashboard/Dashboard_Home.py.

Project Structure

.
├── README.md
├── dashboard/
│   ├── Dashboard_Home.py       # Main entry point
│   ├── eda_utils.py            # Shared EDA utilities
│   ├── requirements.txt
│   ├── pages/                  # One page per dataset
│   └── data/                   # CSV datasets (gzip compressed)
└── reports/                    # Generated markdown reports

Insights Highlights

  • Bank Marketing: Conversion rate ~11%, key drivers are last contact duration and Euribor 3-month rate
  • Credit Card Fraud: Only 0.17% of transactions are fraudulent — extreme class imbalance
  • Loan Default: Grade G loans default at 3x the rate of Grade A loans
  • Medical Insurance: Smokers pay 3-4x more in premiums than non-smokers
  • NYC Taxi: 72% of trips are paid by credit card; average tip is $3.62
  • Online Retail: UK accounts for 80%+ of revenue; October-November are peak months
  • Bank Stocks: JPM and C have the highest intraday volatility among the 9 tickers
  • Steam Games: Free-to-play games dominate player counts; price correlates weakly with rating
  • Video Game Sales: North America leads at ~41% of global sales; PS2 is the best-selling platform

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages