In stix-global/eutradeflows: Download bilateral trade flow data from Eurostat Comext

knitr::opts_chunk$set(echo = FALSE)
knitr::opts_knit$set(root.dir="../..") # file paths are relative to the root of the project directory
library(eutradeflows)
library(dplyr)
library(tidyr)
library(ggplot2)

Introduction

The purpose of this document is to select the main comodities under each CN2 chapter.

Load data

Load data from rds files (Serialization Interface for R Objects). See the notebook bio_imports_update_cache.Rmd to update the cached data.

# Load metadata
partner_names <- readRDS("data_raw/env_impact/partner_names.rds")
reporter_names <- readRDS("data_raw/env_impact/reporter_names.rds")
product_names <- readRDS("data_raw/env_impact/product_names.rds")

product_names_HS2 <- product_names %>% 
    mutate(productcode2d = substr(productcode,1,2)) %>% 
    filter(productcode == productcode2d)

# Load data from cache
system.time(
    yearly <- readRDS("data_raw/env_impact/extra_eu_bio_imports.rds")
)
# 15 seconds for 63 countries
#   user  system elapsed 
# 14.999   0.382  15.389 

yearly <- yearly %>% 
    mutate(year = period %/% 100,
           productcode2d = substr(productcode, 1,2)) 

Aggregate

main_products <- yearly %>% 
    filter(year > max(year) - 5) %>% 
    # Calculate the sum for each CN8 product code 
    group_by(year, productcode2d, productcode) %>% 
    summarise(tradevalue = sum(tradevalue),
              weight = sum(weight),
              tradevalue_b = round(sum(tradevalue)/1e9,1)) %>% 
    # Calculate the percentage of trade within each CN2 product group and each year
    group_by(productcode2d, year) %>% 
    arrange(productcode2d, year, desc(tradevalue))  %>% 
    mutate(sumtradevalue = sum(tradevalue, na.rm = TRUE),
           tradevalue_pct = round(tradevalue / sumtradevalue,2),
           tradevalue_pct_cumul = cumsum(tradevalue_pct)) %>% 
    # Keep only products representing 80% of the total trade value
    filter(tradevalue_pct_cumul <= 0.8 | tradevalue_pct >= 0.8) %>% 
    left_join(product_names, by="productcode") %>%
    mutate(productdescription = substr(productdescription, 1, 20)) 

\pagebreak

Main products at 8 digit level

List of products which represent at least 60% of total trade under the given CN2 chapter (group).

for (this_product2d in unique(yearly$productcode2d)){
    # if(this_product2d=="03") break # Short loop for development
    # Sub section title 
    product_description <- product_names_HS2$productdescription[product_names_HS2$productcode2d==this_product2d]
    cat(sprintf('\n\n## %s %s \n\n',this_product2d, substr(product_description,1,63)))
    # Filter only products under this CN2 code
    df <- main_products %>% 
        filter(productcode2d == this_product2d) 
    # Display 
    df %>% 
        ungroup() %>% 
        mutate(year = ifelse(year==lag(year, default=0),"",year)) %>% 
        select(year, productcode, productdescription, tradevalue, tradevalue_pct, weight) %>% 
        knitr::kable() %>% 
        print()
    # Show full description
    product_names %>% 
        filter(productcode %in% unique(df$productcode)) %>% 
        knitr::kable(format = 'latex') %>% 
        kableExtra::column_spec(2, width = "50em") %>% 
        print()
}

# Encoding issue
# ! Package inputenc Error: Unicode character ├ (U+251C)
# (inputenc)                not set up for use with LaTeX.
# l.10836 ..., N.E.S. (EXCL. IN LIQUID FORM AT 20├
#                                                   É┬░C)\\ 
x <- product_names$productdescription[grepl("├",product_names$productdescription)]
bli <- gregexpr(pattern ='2',x)
bla <- sapply(bli, `[[`, 1)
min(bla[bla>3])
# The issue is not only with this special character but also with those that follow 
# Truncate those to 40 characters
y <- ifelse(grepl("├",product_names$productdescription),
       substr(product_names$productdescription,1,40),
       product_names$productdescription)
summary(nchar(y))

stix-global/eutradeflows documentation built on Nov. 13, 2020, 9:23 p.m.