DataOdyssey

DataOdyssey is an outcome-focused analytics portfolio that turns Python, R, and Jupyter analysis projects into clear data stories. The app is built around what each project reveals: health burden, housing price signals, geopolitical risk concentration, fitness performance patterns, and trading downside exposure.

The goal is not to display notebooks as raw files. DataOdyssey brings the results forward first, then connects each result back to the charts, notebooks, reports, and source material that produced it.

Data Story

DataOdyssey follows a simple narrative structure:

• Start with the decision or question the data can help answer
• Show the strongest outcome from the analysis
• Support the outcome with evidence charts and metrics
• Preserve the original notebooks and reports for deeper inspection

This makes the portfolio easier to read like a set of case studies rather than a folder of disconnected analysis files.

Case Studies

Project	Data Story	Source Material
COVID-19 Socioeconomic Analysis	Connects country-level pandemic burden with mortality, geography, continent mapping, and socioeconomic context.	Johns Hopkins CSSE COVID-19 time-series data, DataHub country/continent codes, and World Bank indicators processed through the COVID19/features pipeline.
Cracow Flat Price Explorer	Explains listing price differences through area, rooms, district, seller type, parking, and neighborhood-level comparisons.	Scraped property listings for sale in Cracow, cleaned from raw_data.csv into a modeling-ready flat-price dataset.
Global Terrorism Trends	Separates high-impact terrorism patterns by fatalities, weapon category, attack type, target type, group activity, and geography.	Global Terrorism Database from START/University of Maryland, with the working copy downloaded from Kaggle and interpreted alongside the GTD codebook.
Polar Fitness Performance Analytics	Links workout duration, calories, time patterns, training type, and heart-rate intensity to personal performance outcomes.	Polar workout export artifacts preserved in the project notebook, PDF report, model output text, and chart gallery.
Trading Results Risk Review	Reviews trading outcomes through profit distribution, drawdown, probability of loss, and simulated downside scenarios.	Trading-results notebook, PDF report, and generated risk charts preserved as the analysis record.

What The App Highlights

• Outcome-first case studies across health, real estate, geopolitics, fitness, and finance
• Searchable project cards with concise evidence and result summaries
• Enhanced result view for the selected case study
• Evidence gallery of generated charts from the original analysis work
• Dark mode and responsive layout for clean reading on desktop and mobile

Data Sources And Provenance

• COVID-19: The workflow pulls Johns Hopkins CSSE global time-series files for confirmed, recovered, and death counts. It adds country and continent metadata from DataHub, then joins World Bank indicators for socioeconomic context.
• Cracow flats: The housing analysis starts from scraped real-estate listing data for properties in Cracow. The wrangling notebook filters the data down to PLN flat listings, cleans location and property fields, and prepares cleaned_data.csv for EDA and modeling.
• Global terrorism: The terrorism story uses the Global Terrorism Database from START/University of Maryland, with the notebook pointing to the GTD codebook and a Kaggle-hosted working copy. The analysis focuses on 2007-2017 incidents and treats missing fatality values conservatively.
• Polar fitness: The fitness project uses personal Polar Flow export data from a Polar watch. The raw JSON export is not included for privacy, but the notebook documents the transformation into workout-level records covering heart rate, calories, duration, sport type, and timing.
• Trading results: The trading analysis uses 92 trades generated by a custom algorithm and exported through an MQL4 script. The private trade dataset is not shared, but the notebook, PDF, and generated charts preserve the methodology and risk review.

The repository intentionally keeps the original notebooks, PDFs, charts, scripts, and model outputs close to the web app. Some raw datasets are referenced or regenerated through scripts rather than fully stored here, especially when the source is external, large, or private.

Tech Stack

• HTML
• Tailwind CSS
• JavaScript
• Python and Jupyter notebooks
• R notebook analysis for terrorism visuals
• Static chart, PDF, and model artifacts

Project Structure

.
|-- index.html          # Main app shell
|-- src/input.css       # Tailwind source and DataOdyssey design layers
|-- styles.css          # Generated production stylesheet
|-- app.js              # Project data, filters, gallery, interactions
|-- DataOdysseyLogo.png # Site logo and favicon source
|-- COVID19/            # COVID feature scripts and notebooks
|-- flats-in-cracow/    # Real estate wrangling, EDA, modeling, and charts
|-- global-terrorism/   # Terrorism analysis notebook and chart gallery
|-- polar/              # Fitness analytics notebook, report, charts, model output
`-- trading-results/    # Trading analysis notebook, report, and risk charts

Run Locally

Install dependencies and build the Tailwind stylesheet:

npm install
npm run build

Then start a local static server:

npm run dev

Open:

http://localhost:3000

Copyright

The DataOdyssey web app design, interface, and implementation additions are copyright (c) 2026 Aarav Sen Mehta. Original analytics notebooks, reports, data artifacts, and inherited source material retain their existing notices and licensing.