Skip to content

This repository contains the data files and the code needed to construct a complete time series of county-to-county migration flows for the period 1998-2015 based on IRS-SOI Migration Data. In addition, I've uploaded also the python scripts I've used to analyse how flows and ties changed after Hurricanes Katrina and Sandy.

Notifications You must be signed in to change notification settings

eugeniopaglino/Hurricanes_and_Migration

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

53 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Hurricanes_and_Migration

This repository contains county-to-county migration data from the Internal Revenue Service (IRS) Statistics of Income (SOI) division. I've assembled two general outflows and inflows files containing county-to-county migration flows for the period 1998-2015, there are the outflow_csv and the inflow.csv files in the outflows and inflows folders respectively. Starting from these aggregated files, I build inflows and outflows matrices for each year in the period 1998-2015. The are organised with each row indicating the sending county for outflows and the receiving county for inflows and each row the destination for outflows and the origin for inflows. These matrices can be found in the outflows and inflows folders, respectively. The IRS releases migration files each year covering a two year period such as 2015-2016. Given the methodology adopted by the IRS, for the generic year1-year2 file, the migration flows mostly refer to movements in year1. Following this convention, I've named the matrices with a two-digits code for year1 and year2 followed by 'in' for inflows matrices and 'out' for outflows matrices. For example, the inflow matrix for the 2001-2002 period is named 0101in.csv.

In addition to the datafiles, I've also uploaded five Python notebooks which I have used to analyse the impact of Hurricanes Katrina and Sandy on the migration system of the affected counties. There are, for each Hurricane, an inflow analysis and an outflow analysis notebook. To have more details on the analysis, refer to the article uploaded in the articles folder. Finally, I've added the Python notebook I've used to construct the dataset I then employed in the regression analysis (refer again to the article).

I've added also the Stata do-file I've used to perform the regression analysis on Katrina and Sandy. In the scripts forlder there are now both a do-file regression_analysis.do and a notebook version of it regression_analysis.ipynb.

I've added the hurricanes.json file to the county_groups folder. This file contains a list of all the hurricanes that hit the US in the 1990-2018 period such that the Federal Emergency Management Agency (FEMA) granted individual assistance to at least one county. For every such hurricane the file containes the name and fip codes of all counties which received individual assistance divided by state. For those hurricanes for which the information was available, I also reported the total cost of the hurricane, the number of deaths it caused, when it started, and when it ended (using the dataset available here https://www.ncdc.noaa.gov/billions/events/US/1980-2019). Given that the FEMA website is not very user-friendly and does not provide accurate search options, I might have overlooked some event which instead should be included. Feel free to point them out.

About

This repository contains the data files and the code needed to construct a complete time series of county-to-county migration flows for the period 1998-2015 based on IRS-SOI Migration Data. In addition, I've uploaded also the python scripts I've used to analyse how flows and ties changed after Hurricanes Katrina and Sandy.

Topics

Resources

Stars

Watchers

Forks

Packages

No packages published