rhdx is an R client for the Humanitarian Exchange Data platform.

Introduction

The Humanitarian Data Exchange platform is the open platform to easily find and analyze humanitarian data.

Installation

This package is not on yet on CRAN and to install it, you will need the remotes package. You can get rhdx from Gitlab or Github (mirror)

## install.packages("remotes")
remotes::install_gitlab("dickoa/rhdx")
remotes::install_github("dickoa/rhdx")

rhdx: A quick tutorial

library("rhdx")

The first step is usually to connect to HDX using the set_rhdx_config function and check the config using get_rhdx_config

set_rhdx_config(hdx_site = "prod")
get_rhdx_config()
## <HDX Configuration>
##   HDX site: prod
##   HDX site url: https://data.humdata.org/
##   HDX API key:

Now that we are connected to HDX, we can search for dataset using search_datasets, access resources withini the dataset page with the get_resources function and finally read the data directly into the R session using read_resource. magrittr pipes operator are also supported

library(tidyverse)
search_datasets("ACLED Mali", rows = 2) %>% ## search dataset in HDX, limit the results to two rows
  pluck(1) %>% ## select the first dataset
  get_resource(1) %>% ## pick the first resource
  read_resource() ## read this HXLated data into R
## # A tibble: 2,516 x 30
##    data_id   iso event_id_cnty event_id_no_cnty event_date  year
##  *   <dbl> <dbl> <chr>                    <dbl> <date>     <dbl>
##  1 2942561   466 MLI2605                   2605 2019-01-26  2019
##  2 2942562   466 MLI2606                   2606 2019-01-26  2019
##  3 2942557   466 MLI2601                   2601 2019-01-25  2019
##  4 2942558   466 MLI2602                   2602 2019-01-25  2019
##  5 2942559   466 MLI2603                   2603 2019-01-25  2019
##  6 2942560   466 MLI2604                   2604 2019-01-25  2019
##  7 2942555   466 MLI2599                   2599 2019-01-24  2019
##  8 2942556   466 MLI2600                   2600 2019-01-24  2019
##  9 2942553   466 MLI2597                   2597 2019-01-23  2019
## 10 2942554   466 MLI2598                   2598 2019-01-23  2019
## # … with 2,506 more rows, and 24 more variables:
## #   time_precision <dbl>, event_type <chr>, actor1 <chr>,
## #   assoc_actor_1 <chr>, inter1 <dbl>, actor2 <chr>,
## #   assoc_actor_2 <chr>, inter2 <dbl>, interaction <dbl>,
## #   region <chr>, country <chr>, admin1 <chr>, admin2 <chr>,
## #   admin3 <chr>, location <chr>, latitude <dbl>,
## #   longitude <dbl>, geo_precision <dbl>, source <chr>,
## #   source_scale <chr>, notes <chr>, fatalities <dbl>,
## #   timestamp <dbl>, iso3 <chr>

read_resource will not work with resources in HDX, so far the following format are supported: csv, xlsx, xls, json, geojson, zipped shapefile, kmz, zipped geodatabase and zipped geopackage. I will consider adding more data types in the future, feel free to file an issue if it doesn’t work as expected or you want to add a support for a format.

Reading dataset directly

We can also use pull_dataset to directly read and access a dataset object.

pull_dataset("acled-data-for-mali") %>%
  get_resource(1) %>%
  read_resource()
## # A tibble: 3,990 x 31
##    data_id   iso event_id_cnty event_id_no_cnty event_date  year
##      <dbl> <dbl> <chr>                    <dbl> <date>     <dbl>
##  1 7173324   466 MLI4111                   4111 2020-07-31  2020
##  2 7173322   466 MLI4109                   4109 2020-07-29  2020
##  3 7173323   466 MLI4110                   4110 2020-07-29  2020
##  4 7173423   466 MLI4107                   4107 2020-07-28  2020
##  5 7173761   466 MLI4108                   4108 2020-07-28  2020
##  6 7173702   466 MLI4104                   4104 2020-07-27  2020
##  7 7173732   466 MLI4103                   4103 2020-07-27  2020
##  8 7173319   466 MLI4102                   4102 2020-07-27  2020
##  9 7173320   466 MLI4105                   4105 2020-07-27  2020
## 10 7173321   466 MLI4106                   4106 2020-07-27  2020
## # … with 3,980 more rows, and 25 more variables:
## #   time_precision <dbl>, event_type <chr>,
## #   sub_event_type <chr>, actor1 <chr>, assoc_actor_1 <chr>,
## #   inter1 <dbl>, actor2 <chr>, assoc_actor_2 <chr>,
## #   inter2 <dbl>, interaction <dbl>, region <chr>,
## #   country <chr>, admin1 <chr>, admin2 <chr>, admin3 <chr>,
## #   location <chr>, latitude <dbl>, longitude <dbl>,
## #   geo_precision <dbl>, source <chr>, source_scale <chr>,
## #   notes <chr>, fatalities <dbl>, timestamp <dbl>, iso3 <chr>

A step by step tutorial to getting data from rhdx

Connect to a server

In order to connect to HDX, we can use the set_rhdx_config function

set_rhdx_config(hdx_site = "prod")

Search datasets

Once a server is chosen, we can now search from dataset using the search_datasets In this case we will limit just to two results (rows parameter).

list_of_ds <- search_datasets("displaced Nigeria", rows = 2)
list_of_ds
## [[1]]
## <HDX Dataset> 4fbc627d-ff64-4bf6-8a49-59904eae15bb
##   Title: Nigeria - Internally displaced persons - IDPs
##   Name: idmc-idp-data-for-nigeria
##   Date: 01/01/2009-12/31/2016
##   Tags (up to 5): displacement, idmc, population
##   Locations (up to 5): nga
##   Resources (up to 5): displacement_data, conflict_data, disaster_data

## [[2]]
## <HDX Dataset> 4adf7874-ae01-46fd-a442-5fc6b3c9dff1
##   Title: Nigeria Baseline Assessment Data [IOM DTM]
##   Name: nigeria-baseline-data-iom-dtm
##   Date: 01/31/2018
##   Tags (up to 5): adamawa, assessment, baseline-data, baseline-dtm, bauchi
##   Locations (up to 5): nga
##   Resources (up to 5): DTM Nigeria Baseline Assessment Round 21, DTM Nigeria Baseline Assessment Round 20, DTM Nigeria Baseline Assessment Round 19, DTM Nigeria Baseline Assessment Round 18, DTM Nigeria Baseline Assessment Round 17

Choose the dataset you want to manipulate in R, in this case we will take the first one.

The result of search_datasets is a list of HDX datasets, you can manipulate this list like any other list in R. We can use purrr::pluck to select the element we want in our list, here it is the first.

ds <- pluck(list_of_ds, 1)
ds
## <HDX Dataset> 4fbc627d-ff64-4bf6-8a49-59904eae15bb
##   Title: Nigeria - Internally displaced persons - IDPs
##   Name: idmc-idp-data-for-nigeria
##   Date: 01/01/2009-12/31/2016
##   Tags (up to 5): displacement, idmc, population
##   Locations (up to 5): nga
##   Resources (up to 5): displacement_data, conflict_data, disaster_data

List all resources in the dataset

With our dataset, the next step is to list all the resources. If you are not familiar with CKAN terminology, resources refer to the actual files shared in a dataset page and you can download. Each dataset page contains one or more resources.

get_resources(ds)
## [[1]]
## <HDX Resource> f57be018-116e-4dd9-a7ab-8002e7627f36
##   Name: displacement_data
##   Description: Internally displaced persons - IDPs (new displacement associated with conflict and violence)
##   Size:
##   Format: JSON

## [[2]]
## <HDX Resource> 6261856c-afb9-4746-b340-9cf531cbd38f
##   Name: conflict_data
##   Description: Internally displaced persons - IDPs (people displaced by conflict and violence)
##   Size:
##   Format: JSON

## [[3]]
## <HDX Resource> b8ff1f4b-105c-4a6c-bf54-a543a486ab7e
##   Name: disaster_data
##   Description: Internally displaced persons - IDPs (new displacement associated with disasters)
##   Size:
##   Format: JSON

Choose a resource we need to download/read

For this example, we are looking for the displacement data and it’s the first resource in the dataset page. We can use pluck on the list of resources or the helper function get_resource(resource, resource_index) to select the resource we want to use. The selected resource can be then downloaded and store for further use or directly read into your R session using the read_resource function. The resource is a json file and it can be read directly using jsonlite package, we added a simplify_json option to get a vector or a data.frame when possible instead of a list.

idp_nga_rs <- get_resource(ds, 1)
idp_nga_df <- read_resource(idp_nga_rs, simplify_json = TRUE, download_folder = tempdir())
idp_nga_df
## # A tibble: 11 x 7
##    ISO3  Name   Year `Conflict Stock… `Conflict New D…
##    <chr> <chr> <dbl>            <dbl>            <dbl>
##  1 NGA   Nige…  2009               NA             5000
##  2 NGA   Nige…  2010               NA             5000
##  3 NGA   Nige…  2011               NA            65000
##  4 NGA   Nige…  2012               NA            63000
##  5 NGA   Nige…  2013          3300000           471000
##  6 NGA   Nige…  2014          1075000           975000
##  7 NGA   Nige…  2015          2096000           737000
##  8 NGA   Nige…  2016          1955000           501000
##  9 NGA   Nige…  2017          1707000           279000
## 10 NGA   Nige…  2018          2216000           541000
## 11 NGA   Nige…  2019          2583000           248000
## # … with 2 more variables: `Disaster New Displacements` <dbl>,
## #   `Disaster Stock Displacement` <dbl>

Using magrittr pipe

All these operations can be chained using pipes %>% and allow for a powerful grammar to easily get humanitarian data in R.

library(tidyverse)

set_rhdx_config(hdx_site = "prod")

idp_nga_df <-
  search_datasets("displaced Nigeria", rows = 2) %>%
  pluck(1) %>%
  get_resource(1) %>% ## get the first resource
  read_resource(simplify_json = TRUE, download_folder = tempdir()) ## the file will be downloaded in a temporary directory

idp_nga_df
## # A tibble: 11 x 7
##    ISO3  Name   Year `Conflict Stock… `Conflict New D…
##    <chr> <chr> <dbl>            <dbl>            <dbl>
##  1 NGA   Nige…  2009               NA             5000
##  2 NGA   Nige…  2010               NA             5000
##  3 NGA   Nige…  2011               NA            65000
##  4 NGA   Nige…  2012               NA            63000
##  5 NGA   Nige…  2013          3300000           471000
##  6 NGA   Nige…  2014          1075000           975000
##  7 NGA   Nige…  2015          2096000           737000
##  8 NGA   Nige…  2016          1955000           501000
##  9 NGA   Nige…  2017          1707000           279000
## 10 NGA   Nige…  2018          2216000           541000
## 11 NGA   Nige…  2019          2583000           248000
## # … with 2 more variables: `Disaster New Displacements` <dbl>,
## #   `Disaster Stock Displacement` <dbl>

Meta