Climate Data Visualization and Analysis Tool

Overview

This project is a Flask-based web application for interactive comparison of climate data between cities.
It loads gridded climate datasets (NetCDF .nc files), computes a range of statistical metrics, and renders interactive visualizations (trends, anomalies, variability, and more) directly in the browser.

Features

• City comparison: Select two cities and compare their climate using historical data.
• 30‑year climatology: Focus on the most recent 30 years to analyse modern climate conditions.
• Trend analysis: Linear warming trends and long‑term temperature evolution.
• Seasonal cycle & monthly means: Monthly climatology plots for both cities.
• Variability metrics: Hourly, daily, monthly, and yearly standard deviations.
• Anomaly distributions: Probability distributions of temperature anomalies.
• Frequency analysis: Power spectrum of temperature time series (dominant periods and cycles).

Tech Stack

• Backend: Python, Flask
• Data & analysis: xarray, NumPy, pandas, SciPy, xarray, SciPy signal processing
• Visualization: Matplotlib, Seaborn (rendered server-side to PNG, sent as base64 to the frontend)
• Frontend: HTML/CSS/JS templates (see templates/index.html)

Project Structure (key files)

• app.py – main Flask application, routes, and all analysis/plotting functions.
• wsgi.py – optional entrypoint if you deploy with a WSGI server (e.g. gunicorn).
• templates/index.html – main UI for selecting cities and displaying results.
• content/ – folder containing climate data files for each city (<city_name>.nc).
• screens/ – screenshots of the UI for quick preview.
• requirements.txt – Python dependencies.
• .gitignore – Git ignore rules for the repo.

Requirements

• Python 3.9+ (3.12 recommended)
• pip or pipenv/poetry for installing dependencies

Install dependencies (with NumPy pinned below 2 to avoid binary-compat issues):

python -m venv .venv
source .venv/bin/activate  # macOS / Linux
pip install -r requirements.txt

Data Preparation

1. Create a folder named content/ in the project root if it does not already exist.

2. Place one NetCDF file per city in this folder, with filenames of the form:

   • content/<city_name>.nc

3. Each file should at least contain:

   • A time dimension
   • A t2m variable (2‑meter air temperature) in Kelvin

The application will automatically list all available cities by scanning content/ for .nc files.

Running the Application Locally

From the project root:

export FLASK_APP=app.py
export FLASK_ENV=development  # optional
flask run

Or, depending on how you prefer to run Flask apps and how app.py is structured, you can often run:

python app.py

Then open http://127.0.0.1:5000/ in your browser and select two cities to compare.

Screenshots

If you just want a quick look at what the app does before running it, here are the screenshots:

Home / city selection

Basic stats + trend comparison

Monthly climatology / seasonality

Variability + anomaly analysis

Frequency-domain / power spectrum

Development Notes

• Plots are generated server‑side using Matplotlib/Seaborn and encoded as base64 images for the frontend.
• Most analysis logic lives in helper functions in app.py (trend, variability, anomaly, and frequency analysis).
• When adding new metrics or visualizations, follow the existing pattern:
  1. Implement a helper function in app.py that returns data + optional base64 image.
  2. Call it from the /compare route and include results in the JSON payload.
  3. Update templates/index.html to render the new section.

License

This project is licensed under the MIT License - see the LICENSE file for details.