Data Analysis and Visualization Repository for Climate Co-exposure Study

This repository contains Jupyter notebooks, Python scripts, and related assets used for data analysis and visualization of for Climate Co-exposure Study.

Data

• Space Resolution: Census tract
• Time Span: 2006 - 2020

We incorporate climate exposure data on:

• Temperature
• Wildfire Smoke PM2.5
• Wildfire

Definitions:

• Heat Day: A day is categorized as a heat day when the average daily temperature surpasses the 95th percentile of the summer months (June, July, August, September) for that specific location across all time and it is greater than 90F. It can be denoted as 1 (occurred) or 0 (not occurred) for each day and place.
• Wildfire Day and Wildfire Smoke Days: Days with wildfire and wildfire smoke respectively.
• Hotspot Day: A day is defined as a hotspot day when two or more events (such as a heat day, wildfire day, and smoke PM day) co-occur. For a day to be categorized as a hotspot, the following conditions must be met:
  • Hotspot (denoted as HD for Heat Day, WF for Wildfire, and SM for Smoke PM): HD & WF & SM
  • Representation: Hotspot = 1 & 1 & 1

This helps in identifying days and places that have a high concentration of multiple climate exposure events.

Repository Structure

Jupyter Notebooks

1. 200_merge_data.ipynb: Merging different data sources.
2. 250_get_metrics.ipynb: Retrieving key metrics such as heat day and wildfire day for analysis.
3. 301_hotspots.ipynb: Identify co-exposure hotspots. 301_stats_hotspots.ipynb: Get stats for each of the hotspots.
4. 302_hotspots_agg_plt_svi.ipynb: Aggregated hotspot data with SVI.
5. 303_lineplot_hotspot_trends.ipynb: Line plots depicting hotspot trends.
6. 305_hotspot_profiling.ipynb: Profiling of hotspots.
7. 310_bivariate_plot.ipynb: Bivariate plotting.
8. 320_get_grid_plots.ipynb: Creating grid plots of exposure per year.
9. 401_demographic_pies.ipynb: Pie charts showing demographic distributions.
10. 401_demographics.ipynb: Other charts showing demographic distributions.
11. 402_age-hotspot.ipynb: Age-based hotspot analysis.
12. 403_income-hotspot.ipynb: Income-based hotspot analysis.
13. 500_kmeans_hotspots.ipynb: K-means clustering for hotspots.
14. association_testing.ipynb: Tests for data association.
15. regression.ipynb: Regression analysis.

Python Scripts

• bivariate_plotting.py: Functions for bivariate plotting.
• helper_400.py: Helper functions for demographic analyses.
• plotting.py: General plotting functions.

Directories

• data: Folder containing raw data and shapefiles.
• outputs: Location for storing generated outputs (e.g., CSVs).
• figures: Folder to store generated figures/plots.
• misc: Miscellaneous files.

Additional Files

• requirements.txt: Python package dependencies for this project.

Installation

To run the Jupyter notebooks and Python scripts, you'll need to have the required packages installed. Install them using:

pip install -r requirements.txt

Data Dictionary

Variable	Description
date	Date of the observation
GEOID	Unique identifier for the geographical region (Census Tract)
wfday	Indicator for wildfire day (binary: 1 if wildfire day, 0 otherwise)
heatday	Indicator for heat day (binary: 1 if heat day, 0 otherwise)
smoke_pm_non_zero	Indicator for days with wildfire smoke particulate matter (PM) presence (binary: 1 if PM present, 0 otherwise)
smoke_pm_gt_five	Indicator for days with wildfire smoke PM > 5ug/m^3 (binary: 1 if PM > 5ug/m^3, 0 otherwise)
hw	Indicator for co-occurrence of heatday and wildfire (binary: 1 if both true, 0 otherwise)
_hws	Indicator if any of heatday, wfday, or smoke_pm_non_zero is true (binary: 1 if any true, 0 otherwise)
hs	Indicator for co-occurrence of heatday and smoke_pm_non_zero (binary: 1 if both true, 0 otherwise)
hws	Indicator for co-occurrence of heatday, wfday, and smoke_pm_non_zero (binary: 1 if all true, 0 otherwise)
ws	Indicator for co-occurrence of wfday and smoke_pm_non_zero (binary: 1 if both true, 0 otherwise)
hs5	Indicator for co-occurrence of heatday and smoke_pm_gt_five (binary: 1 if both true, 0 otherwise)
hws5	Indicator for co-occurrence of heatday, wfday, and smoke_pm_gt_five (binary: 1 if all true, 0 otherwise)
ws5	Indicator for co-occurrence of wfday and smoke_pm_gt_five (binary: 1 if both true, 0 otherwise)
heatday_2d	Indicator for heat day occurrence within 2 days (binary: 1 if occurred, 0 otherwise)
smoke_pm_non_zero_2d	Indicator for smoke PM presence within 2 days (binary: 1 if occurred, 0 otherwise)
wfday_2d	Indicator for wildfire day occurrence within 2 days (binary: 1 if occurred, 0 otherwise)
_hws_2d	Indicator if any of heatday, wfday, or smoke_pm_non_zero occurred within 2 days (binary: 1 if any true, 0 otherwise)
hs_2d	Indicator for co-occurrence of heatday and smoke_pm_non_zero within 2 days (binary: 1 if both true, 0 otherwise)
hws_2d	Indicator for co-occurrence of heatday, wfday, and smoke_pm_non_zero within 2 days (binary: 1 if all true, 0 otherwise)
ws_2d	Indicator for co-occurrence of wfday and smoke_pm_non_zero within 2 days (binary: 1 if both true, 0 otherwise)

Notes:

• s5: Represents wildfire smoke with a concentration greater than 5ug/m^3.
• _2d: Denotes that the climate co-exposure event occurred within a span of 2 days.

Contributing

If you'd like to contribute to this repository, please fork the repository and submit a pull request.

Contact

For questions or feedback, reach out to Kate Hu.