Replication Code for Assessing Land Use Change Trajectories following Food Insecurity Shocks in 25 Low- and Middle-Income Countries

Replication code for the analysis performed in Assessing Land Use Change Trajectories following Food Insecurity Shocks in 25 Low- and Middle-Income Countries  

fiExpand.csv provides a long panel dataset (i.e. each unit-time has its own row) to allow for matching and regression analysis. Data has already been extracted and coallated from raster to the second level administrative units, which are the unit of analysis in this study. 

Numbered R scripts walk through the steps needed to run analyses performed for this study, including linear regressions, event studies, and the IV-inspired analysis of food insecurity drivers. 1_Matching reads in fiExpand.csv and uses the MatchIt R package to produce variable matching for all data, and for data split by urbanization level. 2_Regression_checks.R the distribution of the residuals and calculates outliers using leverage statistics. These outliers are saved to csvs, also included above in the outliers_csvs.zip archive. 3_fi_event_study.R runs the event study analyses on the matched datasets and produces the event study plots (Figure 3 and Figure 5). 4_fi_regressions.R runs the regression analyses on the matched datasets and produces the regression coefficient plots (Figure 2 and Figure 4). 5_fi_IVs runs the IV-inspired analysis that shows how the drivers of food insecurity mediate the outcomes of food insecurity shocks and produces results shown in Table 1. Finally, the case studies for Guatemala, Nigeria and Mozambique are all run in separate scripts, with matching and regression analyses included in each script. Then, the combined case study figures (Figures 6 and 7) are produced in combined_casestudy_graph.R. Figure 1 is produced in python in the Figure_1.ipynb. Administrative boundaries for Figure 1 are downloaded from GAUL (https://data.amerigeoss.org/dataset/a6baab0a-66b8-47ca-84a2-3553be80a574) and filtered to countries included in the study.