In paul4forest/tradeflows: Process forest products trade flow data from COMTRADE and COMEXT

library(knitr)
opts_knit$set(root.dir="../..") # file paths are relative to the root of the project directory
library(tradeflows)
library(dplyr)
library(FAOSTAT)

Merge Comtrade reporters with ITTO region definitions based on reportercode and COMTRADE CODE

reporterITTO <- read.csv("data-raw/ITTO_reporters.csv")
names(reporterITTO)[grepl("tropical",names(reporterITTO))] <- "tropical"
reporterITTO <- reporterITTO %>%
    rename(reportercode = COMTRADE.CODE,
           reporter = COMTRADE.COUNTRIES,
           region = Regional.Distribution)
reporterITTO <- reporterITTO %>% 
    merge(select(reportercomtrade, reportercode, reporter),
          by="reportercode",all.y=TRUE)
reporterITTO %>% head(15) %>% kable
unique(reporterITTO$region)

Countries which don't belong to any region

reporterITTO %>% filter(is.na(region)) %>% kable

Merge Comtrade reporters with FAO region definitions based on reportercode and UN_CODE

To be added to the list of reporter comtrade, inspired by work on another package, see GFPMoutput/rawdata/Merge GFPM and FAO regions and country codes 0.1.r

FAOcountryProfile2 <- FAOcountryProfile %>%
    select(reportercodefao = FAOST_CODE, 
           reporteriso = ISO3_CODE,
           reporternamefao = FAO_TABLE_NAME)

regions <- FAOregionProfile %>%
    filter(!is.na(UNSD_MACRO_REG) & !is.na(FAOST_CODE)) %>%
    select(reportercode = UN_CODE,
           reportercodefao = FAOST_CODE,
           region = UNSD_MACRO_REG,
           subregion = UNSD_SUB_REG) %>%
    merge(FAOcountryProfile2, by="reportercodefao") %>%
    arrange(region, subregion)
str(regions)
# str(FAOcountryProfile)
# str(FAOregionProfile)
# Should merge based on the ISO3 code instead
reportercomtrade2 <- select(reportercomtrade,reportercode, reporter) %>%
    merge(regions, by="reportercode", all.x=TRUE) 

GFPM regions

Add a type of region usefull for the analysis from a North American perspective.

createGFPMRegion = function(region, subregion){
    GFPMregion = NULL
    if (region %in%  c("Asia", "Europe", "Africa", "Oceania")){
        GFPMregion = region
    } 
    if (subregion=="South America"){
            GFPMregion = "South America" 
    }
    if (subregion%in%c("Caribbean", "Central America", "Northern America")){
            GFPMregion = "North/Central America"
    }
    return(GFPMregion)
}
# cm$GFPMRegion = createGFPMRegion(cm$UNSD_MACRO_REG, cm$UNSD_SUB_REG)
# FAOregions$GFPM_REG = mapply(createGFPMRegion,
#                              cm$UNSD_MACRO_REG, cm$UNSD_SUB_REG)
regions %>% 
    mutate(GFPMregion = createGFPMRegion(region, subregion)) %>%
    head(10)

Some country names are missing from the regional aggregates

# Some reporters are missing from the database
missingreporter <- reportercomtrade2 %>% filter(is.na(region))

# Some were present in the FAOregionProfile tables :
missingreporter[missingreporter$reportercode %in% FAOregionProfile$UN_CODE,]

# But most were not present in the FAOregionProvile table
# For example France's un code 251 wasn't in the FAO region aggregates
regions %>% filter(reportercode == 251) 
# Therefore the region is missing
reportercomtrade2 %>% filter(reporter=="France") %>% kable
# However France is present in the FAOstat dataset, under a different code
FAOcountryProfile %>% filter(FAO_TABLE_NAME=="France") %>% kable

# Missing countries not present in the FAOregion table
missingreporter %>% filter(!reporter %in% regions$reporternamefao) %>% kable

################################################## #
# Missing countries present in the FAOregion table #
################################################## #
missingreporter2 <- select(reportercomtrade,reportercode,reporter) %>%
    merge(regions, by="reportercode", all.x=TRUE) %>%
    filter(is.na(region)) %>%
    filter(reporter %in% regions$reporternamefao) %>%
    select(reportercode, reporter)
missingreporter2 %>% kable

regions2 <- regions %>%
    filter(reporternamefao %in% missingreporter$reporter) %>%
    merge(missingreporter2, by.x ="reporternamefao", by.y="reporter")
regions2 %>% kable

# Change their reportercode in the region table so that they match
regions2 <- regions2 %>% 
    rename(reportercode = reportercode.y) %>%
    select(-reportercode.x)

# put them back into the regions table
regions <- rbind(regions, regions2)
# recreate reportercomtrade2
reportercomtrade2 <- select(reportercomtrade,reportercode,reporter) %>%
    merge(regions, by="reportercode", all.x=TRUE) %>%
    arrange(region, subregion)

List of country names belonging to each region

for (subr in unique(reportercomtrade2$subregion)){
    countries <- filter(reportercomtrade2, subregion==subr)
    nbcountries <- nrow(countries)
    cat("\n * __", subr, "__: (", 
         nbcountries," countries)\n", sep="")
    cat(countries$reporter,sep=", ")
}

List of country which don't belong to any region

cat(reportercomtrade2$reporter[is.na(reportercomtrade2$region)], sep="| ")

Merge Comtrade reporters with FAO region definitions based on reporteriso and ISO3_CODE

Load the sawnwood dataset and get iso codes from there

load("data-raw/comtrade/4407.RData")
reportercomtrade3 <- dtf %>% 
    renamecolumns()
# There are more partners than reporters, take reporters
length(unique(reportercomtrade3$reporteriso)) 
length(unique(reportercomtrade3$partneriso)) 
# Partners not in reporters
unique(reportercomtrade3$partneriso)[!unique(reportercomtrade3$partneriso) %in%  unique(reportercomtrade3$reporteriso)]
# Reporters not in partners
unique(reportercomtrade3$reporteriso)[!unique(reportercomtrade3$reporteriso) %in% unique(reportercomtrade3$partneriso)]
# 
#merge based on iso code
# get a table of partnercode, partner, partneriso

reportercomtrade3 <- dtf %>% 
    renamecolumns %>%
    select(reportercode = partnercode,
           reporter = partner,  
           reporteriso = partneriso) %>%
    unique
# Those which are not included
reportercomtrade2$reporter[!reportercomtrade2$reporter %in% reportercomtrade3$reporter]

# Merge based on iso3 code
# remove those which don't have an iso code
regions3 <- regions %>%
    filter(!is.na(reporteriso)) %>%
    select(reporteriso, region, subregion)

reportercomtrade3 <- reportercomtrade3 %>%
    merge(regions3, by="reporteriso")

No ISO code for China

# There is an  ISO code for China in the sawnwood dataset
dtf %>% renamecolumns %>% filter(reportercode==156) %>%
    select(reporter, reportercode,  reporteriso) %>% unique
# But there is no ISO code for China code 351 in FAO tables
filter(regions, subregion == "Eastern Asia") %>% kable
# Or there is no region and subregion mentionned for china code 41
filter(regions, reporteriso == "CHN") %>% kable
# China mainland is a country, but it doesn't have any regions mentionned
# filter(FAOcountryProfile, ISO3_CODE == "CHN")
FAOregionProfile %>%
    #filter(FAOST_CODE==351,FAOST_CODE==41,FAOST_CODE==357)
    filter(FAO_TABLE_NAME == "China") %>%
    select(FAOST_CODE, ISO3_CODE, UN_CODE, OFFICIAL_FAO_NAME) %>%
    kable

List of country names belonging to each region

for (subr in unique(reportercomtrade3$subregion)){
    countries <- filter(reportercomtrade3, subregion==subr)
    cat("\n * __", subr, "__: (", 
        nrow(countries) ," countries)\n", sep="")
    cat(countries$reporter,sep=", ")
}

List of country which don't belong to any region

cat(reportercomtrade3$reporter[is.na(reportercomtrade3$region)], sep="| ")
# Some are not in the list sent by comtrade
reportercomtrade$reporter[!reportercomtrade$reporter %in% reportercomtrade3$reporter]
# but all of them are in the sawnwood dataset
reportercomtrade3$reporter[reportercomtrade3$reporter %in% dtf$reporter]
# How about china?
reportercomtrade3 %>% filter(reporter=="China")
# reportercomtrade3 %>% filter(reportercode.x==156 | reportercode.y==156 )
reportercomtrade %>% filter(reporter =="China")

Sawnwood trade quantity weight and value by subregion

swd <- dtf  %>% 
    renamecolumns %>%
    merge(select(reportercomtrade3, reportercode, region, subregion),
          by="reportercode") %>%
    group_by(subregion) %>%
    summarise(quantity = sum(quantity, na.rm=TRUE),
              weight = sum(weight, na.rm=TRUE), 
              tradevalue = sum(tradevalue)) %>%
    mutate(price = tradevalue / quantity) %>%
    arrange(-tradevalue)
kable(swd)
# Same for swd99

Compare reportercomtrade2 and reportercomtrade3

comp2 <- reportercomtrade2 %>% 
    group_by(region, subregion) %>%
    summarise(nbcountries2 = n())
comp3 <- reportercomtrade3 %>% 
    group_by(region, subregion) %>%
    summarise(nbcountries3 = n())
merge(comp2, comp3) %>% 
    mutate(diff = nbcountries2 - nbcountries3) %>%
    arrange(diff) %>%
    kable
# Are there differences ?
rdiff <- merge(reportercomtrade2, reportercomtrade3,
               by = "reporter", all=TRUE)
filter(rdiff, reportercode.x != reportercode.y)

# All reporteriso seem to be identicall
filter(rdiff, reporteriso.x != reporteriso.y)
# However reportercodes are not all identicall

Full table for those areas which don't have a subregion

rdiff %>% 
    select(reporter, reportercode.x, reportercode.y, 
           subregion.x, subregion.y, reporteriso.x) %>%
    filter(is.na(subregion.x) | is.na(subregion.y)) %>%
    kable

paul4forest/tradeflows documentation built on Oct. 8, 2019, 10:35 a.m.