A joint project to explore Big Data, Data Science tools, and visualization in Python. Utilized two datasets to explore wildfires in the United States. The first dataset tracked wildfire occurrances in all 50 states over 14 years with 1.9m records. The second dataset tracked hourly rainfall totals over 5,500 weather stations between 1940 to the present. These datasets were merged using geolocation to study the relationship between rainfall and wildfire rates. Exploratory analyses were done in Python and Tableau.
This project requires you to install Anaconda 3.7 and make use of Jupyter Notebooks for Python. All packages should be installed based on the import statements at the top of the IPYNB file. All Python work is saved in Exploring the wildfires in US.ipynb and final presentation is available within Wildfires_presentation_1MAR9.pptx as well a shortened version on https://youtu.be/kxaEjTdzNXs. A complete version of the project (567mb zipped) can be found here: https://onedrive.live.com/?authkey=%21ABKvLoSz%5F5z%5Fo40&cid=8CAC97F049439608&id=8CAC97F049439608%211574&parId=root&action=locate
Tableau was also used to create some of the visualizations.
The datasets themselves are too large to be stored on github, so they can be found in the acknowledgement sections.
- Anaconda 3.7
- Python Jupyter Notebooks
- Tableau
Daniel Lesser, Sagnik Rana Carnegie Mellon University Big Data and Data Science Spring 2019
- Kaggle competition US wildfire data https://www.kaggle.com/rtatman/188-million-us-wildfires
- National Oceanic and Atmospheric Adminstration U.S. Hourly Precipitation Data https://data.nodc.noaa.gov/cgi-bin/iso?id=gov.noaa.ncdc:C00313