Checking on data issues

This commit is contained in:
Alex Gebben Work 2026-05-12 17:08:55 -06:00
parent f4b3d3c683
commit e8beb0bbbf

23
temp.r
View File

@ -1,14 +1,17 @@
library(tidyverse)
library(fixest)
DF <- readRDS("Data/Output_Data/Full_Data_Set.rds") %>% mutate(year=as.numeric(year)) %>% group_by(wdid) %>% mutate(SBD1=max(SBD1)) %>% ungroup
DF %>% filter(SBD1_year<2011) %>% pull(SBD1_year) %>% unique
DF[DF$SBD1_year==2009,"SBD1_year"] <- 2011
DF[DF$SBD1==1,"POST"] <-
DF$POST <- 0
DF[DF$SBD1==1& DF$SBD1_year>=DF$year,"POST"] <- 1
#DF$POST <- ifelse(DF$year>=2011,1,0)
feols(log(AF+0.01)~SBD1+POST|wdid,data=DF)
DF$SBD1 <- ifelse(DF$SBD1_year!=Inf,1,0)
DF$SBD1_year <- ifelse(DF$SBD1_year==2009,2011,DF$SBD1_year)
DF$POST <- ifelse(DF$year>=DF$SBD1_year,1,0)
feols(AF~SBD1*POST+CREP_temp+CREP_perm+SBD_half_fallow+SBD1_purchase+SBD1_temp_fallow+SB22_028|wdid+SBD2+SBD3+SBD4+SBD5+SBD6+wdid+ditch_2000812^year+ditch_2000631^year+ditch_Other^year+ditch_2000829^year+ditch_2000798^year+ditch_2000816^year+ditch_2000753^year+ditch_2000753^year+ditch_2000623^year+ditch_2200627^year+ditch_2200541^year+ditch_3500570^year,data=DF)
###########
#DF$SBD1_year <- ifelse(DF$SBD1_year==Inf,10000,DF$SBD1_year)
SUN_MOD <- feols(AF~sunab(SBD1_year,year)+CREP_temp+CREP_perm+SBD_half_fallow+SBD1_purchase+SBD1_temp_fallow+SB22_028+per_alfalfa+per_potatoes|SBD1+SBD2+SBD3+SBD4+SBD5+SBD6+ditch_2000812^year+ditch_2000631^year+ditch_Other^year+ditch_2000829^year+ditch_2000798^year+ditch_2000816^year+ditch_2000753^year+ditch_2000753^year+ditch_2000623^year+ditch_2200627^year+ditch_2200541^year+ditch_3500570^year,data=DF)
SUN_MOD <- feols(AF~sunab(SBD1_year,year)|wdid+SBD1+SBD2+SBD3+SBD4+SBD5+SBD6+ditch_2000812^year+ditch_2000631^year+ditch_Other^year+ditch_2000829^year+ditch_2000798^year+ditch_2000816^year+ditch_2000753^year+ditch_2000753^year+ditch_2000623^year+ditch_2200627^year+ditch_2200541^year+ditch_3500570^year,data=DF %>% filter(SBD1_year==2011|SBD1_year==Inf))
etable(SUN_MOD,agg='cohort')
coefplot(SUN_MOD)
#######CHECK COHORTS IN HYDROBASE LOOKS LIKE MISSING DATA IN SOME LATER COHORTS. DATA MAY NEED WORK
GRAPH_DATA <- DF %>% group_by(SBD1_year,year) %>% summarize(AF=mean(AF)/max(AF))
ggplot(GRAPH_DATA,aes(x=year,y=AF,color=as.factor(SBD1_year),group=SBD1_year))+geom_point()+geom_line()