projections and figures

data/gdp_0.csv

@@ -0,0 +1,12 @@
+# Index: region
+# Unit: $million
+
+Region,gdp_0
+EU,1.7276714501653694e7
+US,1.6755755949456695e7
+OT,9.422720364817122e6
+EE,4.163916626016402e6
+CA,1.5042627818163713e7
+IA,1.1780218688261595e7
+AF,7.107101842851756e6
+LA,7.712450726531674e6

data/pop_0.csv

@@ -1,4 +1,5 @@
 # Index: region
+# Unit: million people
 
 region,pop_0
 EU,503.7099111362591

src/project.R

@@ -0,0 +1,81 @@
+## setwd("~/research/iamup/SSP-Extensions")
+
+library(reshape2)
+library(dplyr)
+library(ggplot2)
+
+pop0 <- read.csv("data/pop_0.csv", skip=2)
+gdp0 <- read.csv("data/gdp_0.csv", skip=2)
+
+for (variable in c('pop', 'gdppc')) {
+ if (variable == 'pop') {
+ baseline <- pop0
+ label <- "Population (millions)"
+ toplimit <- 30000
+ } else {
+ baseline <- data.frame(region=gdp0$Region, gdppc0=gdp0$gdp_0 / pop0$pop_0)
+ label <- "GDP per capita (USD)"
+ toplimit <- 30000000
+ }
+ names(baseline)[2] <- 'value0'
+
+ all.ssp.df <- data.frame()
+ for (ssp in 1:5) {
+ ## Add 2020 from original data
+ if (variable == 'pop') {
+ orig <- read.csv(paste0("data/ssp", ssp, "_pop_rate.csv"), skip=2)
+ orig2 <- melt(orig, 'year', variable.name='region', value.name='rate')
+ orig2$rate.lb <- orig2$rate
+ orig2$rate.ub <- orig2$rate
+ } else {
+ orig.gdp <- read.csv(paste0("data/ssp", ssp, "_gdp_rate.csv"), skip=2)
+ orig <- orig.gdp - read.csv(paste0("data/ssp", ssp, "_pop_rate.csv"), skip=2)
+ orig$year <- orig.gdp$year
+ orig2 <- melt(orig, 'year', variable.name='region', value.name='rate')
+ orig2$rate.lb <- orig2$rate
+ orig2$rate.ub <- orig2$rate
+ }
+
+ df <- read.csv(paste0("results/ssp", ssp, "-", variable, ".csv"))
+ df2 <- rbind(subset(orig2, year == 2020), df)
+
+ years <- c(2015, unique(df2$year))
+ last.df <- baseline
+ last.df$value0.lb <- baseline$value0
+ last.df$value0.ub <- baseline$value0
+ all.df <- cbind(year=2015, last.df)
+ for (tt in 2:length(years)) {
+ workdf <- last.df %>% left_join(subset(df2, year == years[tt]))
+ value1 <- workdf$value0 * (1 + workdf$rate / 100)^(years[tt] - years[tt-1])
+ value1.lb <- workdf$value0.lb * (1 + workdf$rate.lb / 100)^(years[tt] - years[tt-1])
+ value1.ub <- workdf$value0.ub * (1 + workdf$rate.ub / 100)^(years[tt] - years[tt-1])
+ last.df <- data.frame(region=workdf$region, value0=value1, value0.lb=value1.lb, value0.ub=value1.ub)
+ all.df <- rbind(all.df, cbind(year=years[tt], last.df))
+ }
+
+ gp <- ggplot(all.df, aes(year, value0, colour=region, fill=region)) +
+ geom_line() + geom_ribbon(aes(ymin=value0.lb, ymax=value0.ub), alpha=.5) +
+ theme_bw() + scale_colour_discrete(name="Region") + scale_fill_discrete(name="Region") +
+ scale_x_continuous(name=NULL, expand=c(0, 0)) + scale_y_continuous(name=label, expand=c(0, 0))
+ ggsave(paste0("figures/ssp", ssp, "-", variable, ".pdf"), gp, width=5, height=5)
+
+ all.ssp.df <- rbind(all.ssp.df, cbind(ssp, all.df))
+
+ names(all.df)[3:5] <- paste0(variable, c('', '.lb', '.ub'))
+ if (variable == 'pop')
+ write.csv(all.df, paste0("results/ssp", ssp, "-", variable, "-million.csv"), row.names=F)
+ else
+ write.csv(all.df, paste0("results/ssp", ssp, "-", variable, "-usd.csv"), row.names=F)
+ }
+
+ gloall.ssp.df <- all.ssp.df %>% group_by(ssp, year) %>% summarize(value=sum(value0), value.lb=sum(value0.lb), value.ub=sum(value0.ub))
+
+ if (variable == 'pop') {
+ gp <- ggplot(gloall.ssp.df, aes(year, value, colour=factor(ssp), fill=factor(ssp))) +
+ geom_line() + geom_ribbon(aes(ymin=value.lb, ymax=value.ub), alpha=.5) +
+ coord_cartesian(ylim=c(0, toplimit)) + theme_bw() +
+ scale_colour_discrete(name="SSP") + scale_fill_discrete(name="SSP") +
+ scale_x_continuous(name=NULL, expand=c(0, 0)) + scale_y_continuous(name=label, expand=c(0, 0))
+ ggsave(paste0("figures/", variable, ".pdf"), gp, width=5, height=5)
+ }
+}