Last updated: 2020-12-08
Checks: 6 1
Knit directory: GeoPKO/
This reproducible R Markdown analysis was created with workflowr (version 1.6.2). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.
The R Markdown file has unstaged changes. To know which version of the R Markdown file created these results, you’ll want to first commit it to the Git repo. If you’re still working on the analysis, you can ignore this warning. When you’re finished, you can run wflow_publish
to commit the R Markdown file and build the HTML.
Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.
The command set.seed(20200629)
was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.
Great job! Recording the operating system, R version, and package versions is critical for reproducibility.
Nice! There were no cached chunks for this analysis, so you can be confident that you successfully produced the results during this run.
Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.
Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.
The results in this page were generated with repository version c194356. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.
Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish
or wflow_git_commit
). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated:
Ignored files:
Ignored: .Rhistory
Ignored: .Rproj.user/
Untracked files:
Untracked: analysis/gallery/
Unstaged changes:
Modified: analysis/gallery.Rmd
Modified: analysis/merge.Rmd
Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.
These are the previous versions of the repository in which changes were made to the R Markdown (analysis/merge.Rmd
) and HTML (docs/merge.html
) files. If you’ve configured a remote Git repository (see ?wflow_git_remote
), click on the hyperlinks in the table below to view the files as they were in that past version.
File | Version | Author | Date | Message |
---|---|---|---|---|
Rmd | 6a20fd9 | Lou van Roozendaal | 2020-12-04 | update global animated map |
html | c0b0ce3 | Nguyen Ha | 2020-12-03 | Build site. |
Rmd | ee18491 | Nguyen Ha | 2020-12-03 | minor edit |
html | a966566 | Nguyen Ha | 2020-12-03 | Build site. |
Rmd | eafa002 | Nguyen Ha | 2020-12-03 | adapting to new dataset version |
Rmd | 5502ba6 | Lou van Roozendaal | 2020-11-26 | UpdateUCDP |
Rmd | c69de3a | Tanushree Rao | 2020-11-25 | minor text updates + removing warning messages |
Rmd | 7cd68d8 | Tanushree Rao | 2020-11-25 | minor edits, removing warning messages |
html | acee4ed | Nguyen Ha | 2020-11-25 | Build site. |
Rmd | b6fea7a | Nguyen Ha | 2020-11-25 | Revised viz, index, incorporated merge attempt 2.rmd |
html | 2011bda | Nguyen Ha | 2020-11-25 | Build site. |
Rmd | 448b94c | Nguyen Ha | 2020-11-25 | Revised viz, index, incorporated merge.rmd |
Rmd | ad652a3 | Tanushree Rao | 2020-11-24 | changed ref to geopko so it opens with v1 data |
Rmd | 6d15379 | Tanushree Rao | 2020-11-24 | updating merge page with new data |
Rmd | 2608fc0 | Tanushree Rao | 2020-09-10 | merge file |
Rmd | 9174a7c | Tanushree Rao | 2020-09-09 | adding kables |
Rmd | 113774e | Tanushree Rao | 2020-09-04 | edits to data |
Rmd | 5afdf38 | Tanushree Rao | 2020-09-03 | adding merge page |
This page shows how to merge Geo-PKO data with conflict data and visualise the results. The examples used here are Uppsala University’s ViEWS project, which forecasts conflict risk, and the Uppsala Conflict Data Programme (UCDP), one of the world’s leading sources of data on armed conflict. Merging these datasets can provide insights into the links between conflict risk and peacekeeping deployments, and help policymakers make effective peacekeeping decisions where the risk of conflict is high.
Load packages.
library(dplyr)
library(sp)
library(tidyr)
library(geojsonio)
library(broom)
library(rgdal)
library(ggplot2)
library(leaflet)
library(sf)
library(spdep)
library(maptools)
library(plyr)
library(rjson)
library(RJSONIO)
library(rmapshaper)
library(htmltools)
library(htmlwidgets)
First, we import the datasets. We’re using the published Geo-PKO dataset, and conflict forecast data from ViEWS for state-based conflict, non-state conflict, and one-sided violence over the next 36 months in Africa. Both datasets offer insights into the sub-national level, using a unique “PRIO-grid” identifier. The PRIO-grid is an innovative geospatial unit from the Peace Research Institute Oslo that divides the world into roughly 100km x 100km squares, allowing geographic analysis beyond the country level to be streamlined.
library(readr)
geopko <- readr::read_csv("data/Geo_PKO_v2.0.csv")
Parsed with column specification:
cols(
.default = col_double(),
source = col_character(),
mission = col_character(),
joined_date = col_character(),
timepoint = col_character(),
location = col_character(),
country = col_character(),
old_xy = col_logical(),
geocomment = col_character(),
comment.on.unit = col_character(),
rpf = col_logical(),
rpf.no = col_logical(),
obs.base = col_logical(),
cantonment = col_logical(),
disarmament = col_logical(),
he.sup.lw = col_logical(),
troop.type = col_character(),
nameoftcc_1 = col_character(),
notroopspertcc_1 = col_character(),
nameoftcc_2 = col_character(),
nameoftcc_3 = col_character()
# ... with 43 more columns
)
See spec(...) for full column specifications.
Warning: 4400 parsing failures.
row col expected actual file
1032 notroopspertcc_4 a double unknown 'data/Geo_PKO_v2.0.csv'
1035 nameoftcc_6 1/0/T/F/TRUE/FALSE Pakistan 'data/Geo_PKO_v2.0.csv'
1035 notroopspertcc_6 1/0/T/F/TRUE/FALSE 150 'data/Geo_PKO_v2.0.csv'
1035 nameoftcc_7 1/0/T/F/TRUE/FALSE Peru 'data/Geo_PKO_v2.0.csv'
1035 notroopspertcc_7 1/0/T/F/TRUE/FALSE 150 'data/Geo_PKO_v2.0.csv'
.... ................ .................. ........ .......................
See problems(...) for more details.
#unzip("data/ViEWS.zip", exdir="data/ViEWS")
predictors <- read.csv("data/ViEWS/ensemble_pgm.csv")
The Geo-PKO dataset includes detail on troop deployment numbers, types of troops, non-troop deployments, and contributing countries.
library(kableExtra)
Attaching package: 'kableExtra'
The following object is masked from 'package:dplyr':
group_rows
kable(geopko[9546:9550,]) %>% kable_styling() %>%
scroll_box(width = "100%", height = "200px")
source | mission | joined_date | timepoint | year | month | location | geosplit | country | latitude | longitude | old_xy | geocomment | zone.de.confidence | battalion | company | platoon | other.size | comment.on.unit | no.troops | rpf | rpf.no | inf | inf.no | fpu | fpu.no | res | res.no | fp | fp.no | eng | sig | trans | riv | he.sup | sf | med | maint | recon | avia | mp | demining | uav | obs.base | cantonment | disarmament | other.type | armor | he.sup.lw | troop.type | no.tcc | nameoftcc_1 | notroopspertcc_1 | nameoftcc_2 | notroopspertcc_2 | nameoftcc_3 | notroopspertcc_3 | nameoftcc_4 | notroopspertcc_4 | nameoftcc_5 | notroopspertcc_5 | nameoftcc_6 | notroopspertcc_6 | nameoftcc_7 | notroopspertcc_7 | nameoftcc_8 | notroopspertcc_8 | nameoftcc_9 | notroopspertcc_9 | nameoftcc_10 | notroopspertcc_10 | nameoftcc_11 | notroopspertcc_11 | nameoftcc_12 | notroopspertcc_12 | nameoftcc_13 | notroopspertcc_13 | nameoftcc_14 | notroopspertcc_14 | nameoftcc_15 | notroopspertcc_15 | nameoftcc_16 | notroopspertcc_16 | nameoftcc_17 | notroopspertcc_17 | unpol.dummy | unmo.dummy | coding quality for UNMO (1=unsure; 0=perfectly fine) | hq | lo | jmco | security.group.dummy | comments | cow_code | gnwo | tcc1 | tcc2 | tcc3 | tcc4 | tcc5 | tcc6 | tcc7 | tcc8 | tcc9 | tcc10 | tcc11 | tcc12 | tcc13 | tcc14 | tcc15 | tcc16 | tcc17 | adm1.id | adm1.name | prioid | iso3c | Month | MonthName |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Map no. 4249 Rev. 08 | UNMIS | 2007: August | 2007 August | 2007 | 8 | Bentiu | 0 | Sudan | 9.233333 | 29.83333 | NA | GNS has two locations pointing to basically the same place. I took the coords for “seat of first-order admin division”, rahter than “populated place”, both are in line with GE. Also, today’s south sudan | NA | 0 | 1 | 0 | 0 | NA | 150 | NA | NA | 1 | 150 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | NA | NA | NA | 0 | 0 | NA | 1 | 1 | India | 150 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 0 | 1 | 0 | 0 | 0 | FALSE | 0 | NA | 625 | 625 | 750 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 622 | Unity | 142980 | SDN | 8 | August |
Map no. 4249 Rev. 08 | UNMIS | 2007: August | 2007 August | 2007 | 8 | Bor | 0 | Sudan | 6.205931 | 31.55633 | NA | GNS has two locations, the “seat of first-order admin division” pointing to location of Google Earth by same name, as well as corresponding to source map, hence that one coded. Also, today’s south sudan | NA | 0 | 0 | 0 | 0 | NA | 0 | NA | NA | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | NA | NA | NA | 0 | 0 | NA | 0 | 0 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 0 | 1 | 0 | 0 | 0 | FALSE | 0 | NA | 625 | 625 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 621 | Jungoli | 138664 | SDN | 8 | August |
Map no. 4249 Rev. 08 | UNMIS | 2007: August | 2007 August | 2007 | 8 | Dilling | 0 | Sudan | 12.050000 | 29.65000 | NA | NA | NA | 0 | 0 | 0 | 0 | NA | 0 | NA | NA | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | NA | NA | NA | 0 | 0 | NA | 0 | 0 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 0 | 1 | 0 | 0 | 0 | FALSE | 0 | NA | 625 | 625 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 720 | South Kordufan | 147300 | SDN | 8 | August |
Map no. 4249 Rev. 08 | UNMIS | 2007: August | 2007 August | 2007 | 8 | Ed Damazin | 0 | Sudan | 11.789100 | 34.35920 | NA | NA | NA | 1 | 4 | 0 | 0 | NA | 1250 | NA | NA | 1 | 650 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | 0 | NA | NA | NA | 0 | 0 | NA | 1, 2, 4, 11, 14 | 1 | Pakistan | 1250 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 0 | 1 | 0 | 2 | 0 | FALSE | 0 | NA | 625 | 625 | 770 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 287 | Blue Nile | 146589 | SDN | 8 | August |
Map no. 4249 Rev. 08 | UNMIS | 2007: August | 2007 August | 2007 | 8 | El Obeid | 0 | Sudan | 13.183333 | 30.21667 | NA | NA | NA | 0 | 1 | 0 | 0 | NA | 150 | NA | NA | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | NA | NA | NA | 1 | 0 | NA | 99 | 0 | unknown | 150 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 0 | 0 | NA | 0 | 0 | FALSE | 0 | “other” troop type refers to LOG company, unknown TCC | 625 | 625 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 721 | North Kordufan | 148741 | SDN | 8 | August |
The predictor database begins with July 2020 and forecasts the risk of conflict over the next 36 months ahead. Here’s a preview of the data within it, showing state-based (sb
), non-state (ns
) and one-sided violence (os
) forecasts. The variable month_id
codes months differently to the Geo-PKO dataset, with every month assigned a different numeric value.
kable(predictors[90545:90550,]) %>% kable_styling() %>%
scroll_box(width = "100%", height = "200px")
month_id | pg_id | average_allwthematic_sb | average_allwthematic_ns | average_allwthematic_os | |
---|---|---|---|---|---|
90545 | 494 | 141560 | 0.0056672 | 0.0040548 | 0.0049985 |
90546 | 494 | 141561 | 0.0054137 | 0.0032145 | 0.0038274 |
90547 | 494 | 141562 | 0.0115811 | 0.0073864 | 0.0045633 |
90548 | 494 | 141563 | 0.0093885 | 0.0043294 | 0.0027875 |
90549 | 494 | 141564 | 0.0085510 | 0.0184678 | 0.0043437 |
90550 | 494 | 141565 | 0.0098269 | 0.0046215 | 0.0037409 |
The Geo-PKO dataset we’re working with includes data from previous years, but for this visualisation, we will only use the latest year (July 2019 - June 2020). First, we need to filter the data to include only that period. This way we’ll be looking at deployment status against projected conflict risk from the end of that period until three years from that period. You’ll see both the Geo-PKO and ViEWS datasets include the PRIO-grid identification variable (pg_id or PRIOID
), which corresponds to a specific grid square on the map. This is what we’ll use to merge the datasets. In addition to filtering for the time period we want, we will also calculate the average number of troops deployed over that time period.
# filtering for troop deployments over Jul 2019 - Jun 2020
geopko2 <- geopko %>%
select(mission, year, month, prioid, no.troops, country, location, latitude, longitude) %>%
mutate_at(vars(longitude, latitude, year, month, no.troops), as.numeric) %>%
filter(year==2019 & month>6 | year==2020 & month<7)
# calculating an average number of troops
geopko3 <- geopko2 %>%
group_by(prioid) %>%
dplyr::mutate(no.troops = mean(no.troops, na.rm=TRUE)) %>%
ungroup() %>%
filter (! duplicated(no.troops)) %>%
mutate(no.troops=round(no.troops))
Here, we also calculate the average conflict risk (state-based, non-state, and one-sided) for each location.
predictors2 <- predictors %>%
group_by(pg_id) %>%
dplyr::mutate(average_allwthematic_sb = mean(average_allwthematic_sb, na.rm=TRUE)) %>%
dplyr::mutate(average_allwthematic_ns = mean(average_allwthematic_ns, na.rm=TRUE)) %>%
dplyr::mutate(average_allwthematic_os = mean(average_allwthematic_os, na.rm=TRUE)) %>%
filter (! duplicated(average_allwthematic_sb))
Finally, we merge the two datasets.
# merging geopko with conflict forecast data
geopko3$pg_id <- geopko3$prioid
priogriddf <- full_join(
geopko3, predictors2,
by = c("pg_id"),
na.rm = TRUE)
Like we mentioned before, the PRIO-grid unit involves dividing the entire world into roughly 100km x 100km squares. That means that if we want to map it, we’ll be working with large files, so keep that in mind when you’re reading in the shapefile:
shapefile <- rgdal::readOGR("data/ViEWS/priogrid.geojson")
OGR data source with driver: GeoJSON
Source: "C:\Users\Nguyen Ha\Documents\DPCR\GeoPKO\Pages\GeoPKO\data\ViEWS\priogrid.geojson", layer: "priogrid"
with 64818 features
It has 1 fields
The shapefile contains both geospatial polygon data and numerical data that corresponds to the ViEWS dataset; specifically, a PRIO-grid ID and a country ID. Here’s what the non-spatial data looks like, showing five rows in the dataset.
kable(shapefile@data[101:106,]) %>% kable_styling() %>%
scroll_box(width = "100%", height = "200px")
x |
---|
54217 |
54218 |
54219 |
54220 |
54221 |
54222 |
To work with the data within this shapefile, we need to fortify the shapefile. We also convert the IDs to rownames to make it easier to work with. And, finally, we merge it with pgnewdf2
, which we created earlier.
# fortify
shapefile@data$id <- rownames(shapefile@data)
shapefile.df <- fortify(shapefile, region = "id")
# merge data of interest
shapefile.df <- merge(shapefile, priogriddf, by.x = "pg_id", by.y = "pg_id", all.x=F, all.y=T, duplicateGeoms=TRUE)
So now shapefile.df
has the new attributes, including variables from both Geo-PKO and ViEWS. id
is a variable that ties the ‘polygon’, or a single square on the grid, to its location on the map
kable(shapefile.df@data[356:360,]) %>% kable_styling() %>%
scroll_box(width = "100%", height = "200px")
pg_id | id | mission | year | month | prioid | no.troops | country | location | latitude | longitude | month_id | average_allwthematic_sb | average_allwthematic_ns | average_allwthematic_os | |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
356 | 89685 | 3245 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 486 | 0.0024289 | 0.0015213 | 0.0013923 |
357 | 89686 | 3246 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 486 | 0.0029618 | 0.0021277 | 0.0020474 |
358 | 89687 | 3247 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 486 | 0.0026179 | 0.0018690 | 0.0020587 |
359 | 89688 | 3248 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 486 | 0.0023497 | 0.0017310 | 0.0018626 |
360 | 89689 | 3249 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 486 | 0.0023957 | 0.0016043 | 0.0016756 |
To map the data, we’re going to use the leaflet
package (and a bunch of others to support it). The first thing we do is set up our colour palette and bins. We’re using the ‘viridis’ colour palette, designed for accessibility and continuous-scale representation. The other thing we include is a small segment of code that fixes spacing between any NA value in the legend, and the remainder of the legend, making it easier to see.
bins <- c(0, 10, 20, 50, 100, 200, 500, 1000, Inf)
pal <- colorNumeric("viridis", NULL)
#to fix spacing of NA in legend
css_fix <- "div.info.legend.leaflet-control br {clear: both;}" # CSS to correct spacing
html_fix <- htmltools::tags$style(type = "text/css", css_fix) # Convert CSS to HTML
Next, let’s map. We include three colour layers to shade squares according to their conflict forecast value. These layers cover state-based conflict, non-state conflict, and one-sided violence. Simple markers show where troops are deployed. Troop deployment numbers are included as labels, which you can see for each square on hover.
map <- leaflet(shapefile.df) %>%
addTiles() %>%
addPolygons(color = "#444444", weight = 0.25, smoothFactor = 0.5,
opacity = 0.05, fillOpacity = 0.4,
fillColor = ~pal(shapefile.df$average_allwthematic_sb),
group = "State-Based Conflict",
highlightOptions = highlightOptions(color = "white", weight = 2,
bringToFront = FALSE)) %>%
addPolygons(color = "#444444", weight = 0.25, smoothFactor = 0.5,
opacity = 0.05, fillOpacity = 0.4,
fillColor = ~pal(shapefile.df$average_allwthematic_ns),
group = "Non-State Conflict",
highlightOptions = highlightOptions(color = "white", weight = 2,
bringToFront = FALSE)) %>%
addPolygons(color = "#444444", weight = 0.25, smoothFactor = 0.5,
opacity = 0.05, fillOpacity = 0.4,
fillColor = ~pal(shapefile.df$average_allwthematic_os),
group = "One-Sided Violence",
highlightOptions = highlightOptions(color = "white", weight = 2,
bringToFront = FALSE)) %>%
addCircleMarkers((data = shapefile.df@data$no.troops>0), lat = ~latitude, lng = ~longitude,
weight = 1, radius = 2, fillOpacity = 0.6, color = "darkblue") %>%
addPolygons(color = "#444444", weight = 0.1, smoothFactor = 0.5,
opacity = 0.0, fillOpacity = 0.0,
fillColor = ~pal(shapefile.df$no.troops),
label=paste("Troops Deployed: ", shapefile.df$No.troops),
labelOptions = labelOptions(
style = list("font-weight" = "normal", padding = "3px 8px", color="blue"),
textsize = "15px", direction = "auto"),
highlightOptions = highlightOptions(color = "white", weight = 2,
bringToFront = FALSE)) %>%
addLegend("bottomright",
pal = pal,
values = shapefile.df$average_allwthematic_sb,
title = "Conflict Forecast",
opacity = 1) %>%
addLayersControl(
baseGroups = c("State-Based Conflict", "Non-State Conflict", "One-Sided Violence"),
options = layersControlOptions(collapsed = FALSE)
)
Warning in validateCoords(lng, lat, funcName): Data contains 10630 rows with
either missing or invalid lat/lon values and will be ignored
map <- map %>% htmlwidgets::prependContent(html_fix) # legend NA fix
# to save as HTML, you can use the following code:
# saveWidget(map, file="pkoviews - priogrid.html")
map
And here we have it: an interactive map to view recent peacekeeping deployments (2019-2020) and projected conflict risk over the next 36 months. Extensions of this visualisation can be even more useful, particularly with a time-slider that can help us identify how the risk of conflict changes given peacekeeping deployments (and vice versa). That might be a project for the future.
Another useful dataset is from the Uppsala Conflict Data Programme, which offers insights into deaths from armed conflict. Looking at this data in conjunction with peacekeeping data can be useful to draw conclusions into peacekeeping given the severity of armed conflict, or lack thereof. How to merge with this dataset, and a few examples, will be added soon. We start with importing “UCDP Georeferenced Event Dataset (GED) Global version 20.1”, of which a small excerpt of 2011 is shown in the table below.
UCDP <- read_csv("data/UCDP/ged201.csv")
Parsed with column specification:
cols(
year = col_double(),
longitude = col_double(),
latitude = col_double(),
where_coordinates = col_character(),
best = col_double(),
deaths_civilians = col_double()
)
kable(UCDP[9546:9550,]) %>% kable_styling() %>%
scroll_box(width = "100%", height = "200px")
year | longitude | latitude | where_coordinates | best | deaths_civilians |
---|---|---|---|---|---|
2019 | 68.23228 | 33.60806 | Dara Qayaq river | 2 | 0 |
2019 | 63.11753 | 31.91940 | Rakin village | 1 | 0 |
2019 | 65.87257 | 36.30936 | Khanaqa village | 7 | 0 |
2019 | 69.42800 | 37.41980 | Qara Tapa village | 17 | 0 |
2019 | 70.21723 | 33.46187 | Babrak Tana ruin | 6 | 6 |
For the GeoPKO data we only used certain variables, and calculate the average troop number by year. The UCDP data has different variables you could use, in this example we took “best”, which indicates “the best (most likely) estimate of total fatalities resulting from integer an event”(see UCDP codebook), and “deaths_civilians”, which tells us how many of this “best” variable were civilian deaths.
# Preparing the GeoPKO dataset
GeoPKO_dataUCDP <- geopko %>%
select(mission, year, location, latitude, longitude, no.troops) %>% #Select only the variables you need
mutate_at(vars(latitude, longitude, no.troops), as.numeric) %>%
group_by(mission, year, location) %>%
mutate(ave.no.troops = as.integer(mean(no.troops, na.rm=TRUE))) %>% #Sum the troop numbers by year through using a combination of group_by & mutate
select(-no.troops) %>% #Deselect the previous troop number variable
distinct() %>% #Delete any duplicate rows
drop_na(ave.no.troops)%>% #Remove NAs from the average troop count
filter(ave.no.troops > 0) # Exclude any troop numbers under the value of 0
# Preparing the UCDP dataset
UCDP_dataframe <- UCDP %>%
select(year, longitude, latitude, where_coordinates, best,deaths_civilians)%>%
drop_na(latitude,longitude) %>%
group_by(year, where_coordinates) %>%
mutate(best = as.integer(mean(best, na.rm=TRUE))) %>% #Take the mean of the "best" variable
mutate(deaths_civilians = as.integer(mean(deaths_civilians, na.rm=TRUE)))%>% #Take the mean of the "deaths_civilians" variable
filter(best > 0 & year>=1995)%>% #Filter that data so that it only keeps rows for the years after 1995.
distinct()
##Setting the aesthetics You can either set all the colours by hand, as seen in pal3, or use the “viridis” package to create a colour scale for you, as shown in pal2. For the UCDP data we used shades of red. The GeoPKO is mapped with viridis, which exists out of blue, green, and yellow.
pal2 <- colorBin((viridis::viridis(10)), GeoPKO_dataUCDP$ave.no.troops, bins = c(1,50,100,500,1000,2000,4000,8000))
pal3 <- colorBin(c("#700524","#8d072e","#981f42","#ed4d3a","#af516c","#d19bab","#dcb4c0"),
UCDP_dataframe$best, bins = c(1,50,100,500,1000,2000,4000,10000,Inf))
Just as we used leaflet for the VIEWS map, we do the same here. Note that in this case we did not merge the two datasets into one dataframe. When hovering over the circles more information on either the UN peacekeeping deployment will be provided or on the UCDP conflict-related deaths. It is possible to use the PRIO-grid ID instead, both datasets include these.
UCDP_Overview_Map <- leaflet() %>%
addTiles(options = providerTileOptions(noWrap = TRUE)) %>%
addMeasure(position = "bottomleft", primaryLengthUnit = "kilometers")%>% #Adds a widget that can measure distances between two places to the map
clearMarkers()%>%
clearShapes()%>%
addLegend("topright",
pal = pal3,
values = UCDP$best,
title = "Fatalities",
opacity = 1) %>%
addLegend("topright",
pal = pal2,
values = GeoPKO_dataUCDP$ave.no.troops,
title = "Peacekeepers",
opacity = 1) %>%
addLayersControl(baseGroups = c("2019", "2018","2015","2010", "2005","2000","1995"),
options = layersControlOptions(collapsed = FALSE), position = "topleft")%>%
addCircleMarkers(data=(GeoPKO2019<-GeoPKO_dataUCDP %>% filter(year==2019)),
color = ~pal2(ave.no.troops), radius = ~(ave.no.troops)^(1/3),
opacity = 0.15, fillOpacity = 0.5,
lng = ~longitude, lat = ~latitude, group = "2019",
label=paste("<strong>UN Peacekeeping Site<br/>Mission:</strong>",GeoPKO2019$mission,
"<br/><strong>Location:</strong>",GeoPKO2019$location,
"<br/><strong>Troops Deployed:</strong>", GeoPKO2019$ave.no.troops)%>%
lapply(htmltools::HTML)) %>%
addCircleMarkers(data = (UCDP2019<-UCDP_dataframe %>% filter(year==2019)),
lng = ~longitude, lat = ~latitude,
color = ~pal3(best),radius = ~(best)^(1/3),
opacity = 0.05, fillOpacity = 0.4,
group = "2019",
label=paste("<strong>UCDP Reported Fatalities<br/>Location:</strong>", UCDP2019$where_coordinates,
"<br/><strong>Total deaths:</strong>", UCDP2019$best,
"<br/><strong>Civilian deaths:</strong>", UCDP2019$deaths_civilians)%>%
lapply(htmltools::HTML)) %>%
addCircleMarkers(data=(GeoPKO2018<-GeoPKO_dataUCDP %>%filter(year==2018)),
color = ~pal2(ave.no.troops),radius = ~(ave.no.troops)^(1/3),
opacity = 0.15, fillOpacity = 0.5,
lng = ~longitude, lat = ~latitude, group = "2018",
label=paste("<strong>UN Peacekeeping Site<br/>Mission:</strong>",GeoPKO2018$mission,
"<br/><strong>Location:</strong>",GeoPKO2018$location,
"<br/><strong>Troops Deployed:</strong>", GeoPKO2018$ave.no.troops)%>%
lapply(htmltools::HTML)) %>%
addCircleMarkers(data = (UCDP2018<-UCDP_dataframe %>%filter(year==2018)),
lng = ~longitude, lat = ~latitude,
color = ~pal3(best),radius = ~(best)^(1/3),
opacity = 0.05, fillOpacity = 0.4,
group = "2018",
label=paste("<strong>UCDP Reported Fatalities<br/>Location:</strong>", UCDP2018$where_coordinates,
"<br/><strong>Total deaths:</strong>", UCDP2018$best,
"<br/><strong>Civilian deaths:</strong>", UCDP2018$deaths_civilians)%>% lapply(htmltools::HTML)) %>%
addCircleMarkers(data=(GeoPKO2015<-GeoPKO_dataUCDP %>%filter(year==2015)),
color = ~pal2(ave.no.troops),radius = ~(ave.no.troops)^(1/3),
opacity = 0.15, fillOpacity = 0.5,
lng = ~longitude, lat = ~latitude, group = "2015",
label=paste("<strong>UN Peacekeeping Site<br/>Mission:</strong>",GeoPKO2015$mission,
"<br/><strong>Location:</strong>",GeoPKO2015$location,
"<br/><strong>Troops Deployed:</strong>", GeoPKO2015$ave.no.troops)%>%
lapply(htmltools::HTML)) %>%
addCircleMarkers(data = (UCDP2015<-UCDP_dataframe %>%filter(year==2015)),
lng = ~longitude, lat = ~latitude,
color = ~pal3(best),radius = ~(best)^(1/3),
opacity = 0.05, fillOpacity = 0.4,
group = "2015",
label=paste("<strong>UCDP Reported Fatalities<br/>Location:</strong>", UCDP2015$where_coordinates,
"<br/><strong>Total deaths:</strong>", UCDP2015$best,
"<br/><strong>Civilian deaths:</strong>", UCDP2015$deaths_civilians)%>% lapply(htmltools::HTML)) %>%
addCircleMarkers(data=(GeoPKO2010<-GeoPKO_dataUCDP %>%filter(year==2010)),
color = ~pal2(ave.no.troops),radius = ~(ave.no.troops)^(1/3),
opacity = 0.15, fillOpacity = 0.5,
lng = ~longitude, lat = ~latitude, group = "2010",
label=paste("<strong>UN Peacekeeping Site<br/>Mission:</strong>",GeoPKO2010$mission,
"<br/><strong>Location:</strong>",GeoPKO2010$location,
"<br/><strong>Troops Deployed:</strong>", GeoPKO2010$ave.no.troops)%>%
lapply(htmltools::HTML)) %>%
addCircleMarkers(data = (UCDP2010<-UCDP_dataframe %>%filter(year==2010)),
lng = ~longitude, lat = ~latitude,
color = ~pal3(best),radius = ~(best)^(1/3),
opacity = 0.05, fillOpacity = 0.4,
group = "2010",
label=paste("<strong>UCDP Reported Fatalities<br/>Location:</strong>", UCDP2010$where_coordinates,
"<br/><strong>Total deaths:</strong>", UCDP2010$best,
"<br/><strong>Civilian deaths:</strong>", UCDP2010$deaths_civilians)%>% lapply(htmltools::HTML)) %>%
addCircleMarkers(data=(GeoPKO2005<-GeoPKO_dataUCDP %>%filter(year==2005)),
color = ~pal2(ave.no.troops),radius = ~(ave.no.troops)^(1/3),
opacity = 0.15, fillOpacity = 0.5,
lng = ~longitude, lat = ~latitude, group = "2005",
label=paste("<strong>UN Peacekeeping Site<br/>Mission:</strong>",GeoPKO2005$mission,
"<br/><strong>Location:</strong>",GeoPKO2005$location,
"<br/><strong>Troops Deployed:</strong>", GeoPKO2005$ave.no.troops)%>%
lapply(htmltools::HTML)) %>%
addCircleMarkers(data = (UCDP2005<-UCDP_dataframe %>%filter(year==2005)),
lng = ~longitude, lat = ~latitude,
color = ~pal3(best),radius = ~(best)^(1/3),
opacity = 0.05, fillOpacity = 0.4,
group = "2005",
label=paste("<strong>UCDP Reported Fatalities<br/>Location:</strong>", UCDP2005$where_coordinates,
"<br/><strong>Total deaths:</strong>", UCDP2005$best,
"<br/><strong>Civilian deaths:</strong>", UCDP2005$deaths_civilians)%>% lapply(htmltools::HTML)) %>%
addCircleMarkers(data=(GeoPKO2000<-GeoPKO_dataUCDP %>%filter(year==2000)), color = ~pal2(ave.no.troops),radius = ~(ave.no.troops)^(1/3),
opacity = 0.15, fillOpacity = 0.5,
lng = ~longitude, lat = ~latitude, group = "2000",
label=paste("<strong>UN Peacekeeping Site<br/>Mission:</strong>",GeoPKO2000$mission,
"<br/><strong>Location:</strong>",GeoPKO2000$location,
"<br/><strong>Troops Deployed:</strong>", GeoPKO2000$ave.no.troops)%>%
lapply(htmltools::HTML)) %>%
addCircleMarkers(data = (UCDP2000<-UCDP_dataframe %>%filter(year==2000)),
lng = ~longitude, lat = ~latitude,
color = ~pal3(best),radius = ~(best)^(1/3),
opacity = 0.05, fillOpacity = 0.4,
group = "2000",
label=paste("<strong>UCDP Reported Fatalities<br/>Location:</strong>", UCDP2000$where_coordinates,
"<br/><strong>Total deaths:</strong>", UCDP2000$best,
"<br/><strong>Civilian deaths:</strong>", UCDP2000$deaths_civilians)%>% lapply(htmltools::HTML)) %>%
addCircleMarkers(data=(GeoPKO1995<-GeoPKO_dataUCDP %>%filter(year==1995)), color = ~pal2(ave.no.troops),radius = ~(ave.no.troops)^(1/3),
lng = ~longitude, lat = ~latitude,
opacity = 0.15, fillOpacity = 0.5,
group = "1995",
label=paste("<strong>UN Peacekeeping Site<br/>Mission:</strong>",GeoPKO1995$mission,
"<br/><strong>Location:</strong>",GeoPKO1995$location,
"<br/><strong>Troops Deployed:</strong>", GeoPKO1995$ave.no.troops)%>%
lapply(htmltools::HTML)) %>%
addCircleMarkers(data = (UCDP1995<-UCDP_dataframe %>%filter(year==1995)),
lng = ~longitude, lat = ~latitude, color = ~pal3(best),radius = ~(best)^(1/3),
opacity = 0.05, fillOpacity = 0.4,
group = "1995",
label=paste("<strong>UCDP Reported Fatalities<br/>Location:</strong>", UCDP1995$where_coordinates,
"<br/><strong>Total deaths:</strong>", UCDP1995$best,
"<br/><strong>Civilian deaths:</strong>", UCDP1995$deaths_civilians)%>% lapply(htmltools::HTML))
# to save as HTML, you can use the following code:
# saveWidget(UCDP_Overview_Map, file="geopko&ucdp - geopko.html")
#Shows the map
UCDP_Overview_Map
And here we have it again: an interactive map to view peacekeeping deployments, in a few selected years, and conflict-related deaths within that same year. We opted to allow the user to select different years to see how both the numbers and locations of both datasets change over time.
If you do want to combine the two datasets into one dataframe there are different ways to do so. As shown with the VIEWS data, you can merge the datasets based on certain variables using “full_join”, causing the two datasets to be combined in one file behind each other (it creates a dataframe in which the variables of one of the dataset occur next to the other), this can be utilized to the UCDP data as well. However, you can also merge datasets “under” each other, for the UCDP we wanted to show this second way of merging. Variables that are the same between the two datasets might go by different names, that is why some of the variable of the UCDP are renamed so that they are in line with the GeoPKO ones. The newly combined dataset will first show all the UCDP rows, and under these the GeoPKO rows will be “pasted”. Merging can be useful for all kinds of reasons, such as statistical analysis.
# Rename variables you want to merge
#UCDP_dataframeMerge = UCDP_dataframe %>% rename (location=where_coordinates)
# Merge the datasets
#Combined_UCDP_GeoPKO<- bind_rows(UCDP_dataframeMerge, GeoPKO_dataUCDP)
sessionInfo()
R version 4.0.2 (2020-06-22)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19041)
Matrix products: default
locale:
[1] LC_COLLATE=English_Sweden.1252 LC_CTYPE=English_Sweden.1252
[3] LC_MONETARY=English_Sweden.1252 LC_NUMERIC=C
[5] LC_TIME=English_Sweden.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] kableExtra_1.1.0 readr_1.3.1 htmlwidgets_1.5.1 htmltools_0.5.0
[5] rmapshaper_0.4.4 RJSONIO_1.3-1.4 rjson_0.2.20 plyr_1.8.6
[9] maptools_1.0-2 spdep_1.1-5 spData_0.3.8 sf_0.9-5
[13] leaflet_2.0.3 ggplot2_3.3.2 rgdal_1.5-16 broom_0.7.0
[17] geojsonio_0.9.2 tidyr_1.1.1 sp_1.4-2 dplyr_1.0.2
loaded via a namespace (and not attached):
[1] nlme_3.1-148 fs_1.5.0 RColorBrewer_1.1-2 webshot_0.5.2
[5] httr_1.4.2 gmodels_2.18.1 rprojroot_1.3-2 tools_4.0.2
[9] backports_1.1.7 R6_2.4.1 KernSmooth_2.23-17 rgeos_0.5-3
[13] DBI_1.1.0 lazyeval_0.2.2 colorspace_1.4-1 raster_3.3-13
[17] withr_2.2.0 gridExtra_2.3 tidyselect_1.1.0 curl_4.3
[21] compiler_4.0.2 git2r_0.27.1 rvest_0.3.6 expm_0.999-5
[25] xml2_1.3.2 scales_1.1.1 classInt_0.4-3 stringr_1.4.0
[29] digest_0.6.25 foreign_0.8-80 rmarkdown_2.3 pkgconfig_2.0.3
[33] highr_0.8 jsonvalidate_1.1.0 rlang_0.4.7 rstudioapi_0.11
[37] httpcode_0.3.0 farver_2.0.3 generics_0.0.2 jsonlite_1.7.1
[41] crosstalk_1.1.0.1 gtools_3.8.2 magrittr_1.5 Matrix_1.2-18
[45] Rcpp_1.0.5 munsell_0.5.0 viridis_0.5.1 lifecycle_0.2.0
[49] stringi_1.4.6 whisker_0.4 yaml_2.2.1 MASS_7.3-52
[53] jqr_1.1.0 grid_4.0.2 gdata_2.18.0 promises_1.1.1
[57] crayon_1.3.4 deldir_0.2-3 lattice_0.20-41 splines_4.0.2
[61] geojson_0.3.4 hms_0.5.3 knitr_1.29 pillar_1.4.6
[65] boot_1.3-25 geojsonlint_0.4.0 codetools_0.2-16 LearnBayes_2.15.1
[69] crul_1.0.0 glue_1.4.1 evaluate_0.14 V8_3.2.0
[73] vctrs_0.3.2 httpuv_1.5.4 gtable_0.3.0 purrr_0.3.4
[77] xfun_0.16 e1071_1.7-3 coda_0.19-4 later_1.1.0.1
[81] viridisLite_0.3.0 class_7.3-17 tibble_3.0.3 workflowr_1.6.2
[85] units_0.6-7 ellipsis_0.3.1