Skip to content

Analyzing and visualizing 200+ years of global climate data using python

License

Notifications You must be signed in to change notification settings

codeananda/geographic_data_analysis

Repository files navigation

🌎 Geographic Data Analysis

Code style: black

A long-term data analysis project completed on Upwork. We worked with 100+ GBs of geospatial data - surface temperature for every 50 km square on the Earth's surface for the last 100 years and several predictions for the next 100 years. We used the Python data science stack throughout (dask, xarray, pandas, numpy, matplotlib, and seaborn among others).

tasmax_max_temp in South-Central Asia

An example plot created for the final deliverable - a journal article for scientific publication.

It was a two-part project that consisted of weekly calls and lead to the following ⭐⭐⭐⭐⭐ reviews:

Screenshot 2021-09-02 at 14 16 12

Screenshot 2021-09-02 at 14 16 03

🤔 The Problem

THE DATASET

We have temperature data for every 50 km^2 cell on the Earth's surface for the last 100 years and predictions for those cells for the next 100 years. The predictions were generated independently by four different laboratories.

THE GOAL

Analyse how the number and intensity of heatwaves is expected to change in the next 100 years. Compare and contrast the predictions made by each of the laboratories.

📕 Noteable Notebooks

  • Check out 4_heatwaves to see how I created 300+ pdfs of plots corresponding to each cell.

  • Check out 5_final_plots to see some of the final deliverables for the project - publication-ready plots of tasmin/tasmax data by cell and by region.

  • A good balance of plots and analysis can be found in the heatwave_analysis.ipynb notebook.

  • The historical_analysis_tasmax.ipynb notebook explores six different methods to define a heatwave and plots their differences for a sample of cells.

📜 Terminology

The dataset consisted of several excel files containing the min/max temperature for every 50 km^2 (cell) of the Earth for the last 100 years. We combined this with four similar datasets containing predictions of the min/max temperature for every 50 km^2 cell of the Earth for the next 100 years. We restricted our analysis just to points on land.

  • Cell - a 50 km^2 grid on Earth given by longitude and latitude point.
  • Tas - Temperature at Surface, this is a particular method geographers and meterologists use to measure the temperature.
  • Tasmin - Temperature at Surface (TAS) minimum value for that day. The minimum temperature recorded on that day for that cell.
  • Tasmax - Temperature at Surface (TAS) maximum value for that day. The maximum temperature recorded on that day for that cell.

🏗 This Repo is a Work-in-Progress

I finished this project in August 2020 and have been kept busy with other projects since then. I am still in the process of getting this into a publication-ready state but I hope this gives you a good idea of my work.

About

Analyzing and visualizing 200+ years of global climate data using python

Resources

License

Stars

Watchers

Forks