Contents of the talk

• Introduction

• API

• R Package

• Dplyr + SQL

• Creating your own API

• Shiny apps

Where to reach me

Twitter and Github: pachamaltese

Email: pacha # dcc * uchile * cl

Introduction

• Open Trade Statistics (OTS) was created with the intention to lower the barrier to working with international economic trade data.

• It includes a public API, a dashboard, and an R package for data retrieval.

Introduction

• Many Latin American Universities have limited or no access to the United Nations Commodity Trade Statistics Database (UN COMTRADE).

• This project shares curated datasets based on UN COMTRADE.

Introduction

The project has a major reproducibility flaw.

Introduction

Hardware and software stack

Introduction

The next three slides are an oversimplification just to explain the work in wide terms.

Introduction

The raw data contains missing flows:

Introduction

Possible solution (Anderson & van Wincoop, 2004 propose 8% rate):

Introduction

Customs have changed their coding systems in order to reflect changes in exported products (i.e. in 1960 nobody exported laptops).

Introduction

• We have 2018 data, similar initiatives have datasets updated to 2016 or 2017.

• But much more important than that, we converted all years to HS rev 2007 to allow time comparisons.

R package

# easy start
install.packages("tradestatistics")

R package

Fiji exports a lot of water. But how much of its exports to the US are actually water?

library(dplyr)
library(tradestatistics)
fji_usa <- ots_create_tidy_data(
  years = 2017, reporters = "fji", partners = "usa",
  include_shortnames = T
)
fji_usa_2 <- fji_usa %>% 
  select(product_shortname_english, export_value_usd) %>% 
  arrange(-export_value_usd) %>% 
  mutate(export_value_share = round(100 * export_value_usd /
                                      sum(export_value_usd, na.rm = T), 2))

R package

fji_usa_2

## # A tibble: 736 x 3
##    product_shortname_english export_value_usd export_value_share
##    <chr>                                <int>              <dbl>
##  1 Water                            233431002              60   
##  2 Processed Fish                    61666883              15.8 
##  3 Non-fillet Fresh Fish             18503975               4.76
##  4 Raw Sugar                         12600118               3.24
##  5 Broadcasting Equipment            10967992               2.82
##  6 Perfume Plants                     7273321               1.87
##  7 Fish Fillets                       5540948               1.42
##  8 Unspecified                        4246687               1.09
##  9 Non-fillet Frozen Fish             4033516               1.04
## 10 Molasses                           3578212               0.92
## # … with 726 more rows

R package

Which country from America is the #1 partner with the European Union (EU-28)?

eu28 <- ots_countries %>% 
  filter(eu28_member == 1) %>% 
  select(country_iso)
ame_eu28 <- ots_create_tidy_data(
  years = 2017, reporters = "c-am", partners = "all",
  table = "yrp"
)

ame_eu28_2 <- ame_eu28 %>% 
  mutate(is_eu28 = ifelse(partner_iso %in% eu28$country_iso, 1, 0)) %>% 
  group_by(reporter_iso, is_eu28) %>% 
  summarise(export_value_usd = sum(export_value_usd, na.rm = T)) %>% 
  group_by(reporter_iso) %>%
  mutate(pct_to_eu28 = export_value_usd / sum(export_value_usd, na.rm = T)) %>% 
  filter(is_eu28 == 1) %>% 
  select(reporter_iso, export_value_usd, pct_to_eu28) %>% 
  arrange(-export_value_usd)

ame_eu28_2

## # A tibble: 48 x 3
## # Groups:   reporter_iso [48]
##    reporter_iso export_value_usd pct_to_eu28
##    <chr>                   <dbl>       <dbl>
##  1 usa              406704170165      0.218 
##  2 can               51259679792      0.109 
##  3 bra               45042481104      0.174 
##  4 mex               34405957170      0.0711
##  5 chl               10937519973      0.144 
##  6 arg               10872575385      0.165 
##  7 per                8375757038      0.170 
##  8 col                8124020268      0.168 
##  9 cri                3922634462      0.246 
## 10 ecu                3917486407      0.169 
## # … with 38 more rows

Code and documentation

github.com/tradestatistics

docs.ropensci.org/tradestatistics

tradestatistics.io

Dplyr + SQL

Your turn

Creating your own API

Your turn

Shiny apps

Your turn

Acknowledgements

• rOpenSci <3: Amanda, Emily, Jorge, Maelle, Mark and Stefanie
• DigitalOcean: Danny
• Highcharter/Design: Joshua and Erasmo