Skip to content

undergraduate capstone project completed by Joseph Kassab, Zeyu Jin, and Yechan Seo under the guidance of Dr. Maura Allaire.

Notifications You must be signed in to change notification settings

Cosmos0603/hurricane_project

Repository files navigation

County-Level Analysis of Hurricane Impacts on Consumption Patterns Across the U.S.

Datasets

FEMA National Risk Index Dataset:

  • provides county level statistics and scores on social vulnerability, hurricane records, hurricane exposure, and hurricane risk
  • publically available at https://arcg.is/0KD9Tq

Nielsen Scanner Consumption Dataset:

  • provides weekly product sales by county across all states
  • not publically available

Census Bureau Dataset:

Hurricane Marker CSV:

  • provides the markers for past hurricane occurrences
  • not publically available

Notebooks Description

  • exploredata_FEMA: explore and preprocess the FEMA data
  • exploredata_psid: explore and preprocess the psid data
  • exploredata_Nielsen: explore and preprocess the Nielsen data
  • exploredata_Census: explore and preprocess the Census data (demo info)
  • explore_visualize_data.ipynb: Data Visualizations
  • hurricane_marker: explore and preprocess the hurricane_marker.csv
  • merge_plot:
    • merge all the preprocessed data into one single data frames
    • visualize the resulted data frame
  • baseline:
    • prepare the data for model training
    • experiment with different baseline models
    • perform county filtering and product grouping
  • Model Evaluation:
    • can be run with the sample data (df_tx_sample.pkl)
    • parameter tuning with randomized search
    • evaluating different models with different feature combinations

About

undergraduate capstone project completed by Joseph Kassab, Zeyu Jin, and Yechan Seo under the guidance of Dr. Maura Allaire.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published