Lazy labeling: regular expressions

AKA how start getting around using grep, grepl, sub, gsub, regexp, gregexp.

Main difference between them is that while grep and grepl matches to pattern within the character vector, regexpr and gregexpr return more detail. sub and gsub perform replacement on the basis of a regular expression.

This is my version of lazy labeling: run the same code changing just the input dataset

file <- "BEN_Benin_2013" #myfile
clabel <- substring(file, 1, 3) #country label is the iso3 code
cname <-sub( ".*_(.*)_.*", "\\1", file) #country name
year <- sub( ".*_(.*)_*", "\\1", file) #census year

More to follow on grep and grepl…

See you at PAA 2018 in Denver!

As PAA 2018 is approaching, I’m looking forward to talk more about spatial prediction and subnational fertility disaggregation using DHS household survey data and our research on fine grid scale mapping of demographic indicators in low income countries at WorldPop (University of Southampton) and Flowminder.

Come meet me at the session “Adolescent Transitions into Parenthood”, on Saturday 28th of April from 11:15am-12:45pm in Plaza Ballroom E (Sheraton Denver Downtown), I will present our paper “The geography of fertility rates in low and middle-income countries: analysis of cross-sectional surveys from 69 countries”.

 

 

Kuwait population pyramids: changes between 2000-2020

Plot made in ggplot2 using geofacet, personal elaboration of Census data.

KWT_pyr2000-2020

Here is the the geofacet gris, although maybe not one of the most exciting ones…

mygrid <- data.frame(
row = c(1, 1, 1, 2, 2, 3),
col = c(1, 2, 3, 2, 3, 3),
code = c("KWT1", "KWT2", "KWT3", "KWT4", "KWT5", "KWT6"),
name = c("Al-Jahra Governorate", "Capital (Al Asimah) Governorate", "Hawalli Governorate", "Al-Farwaniya Governorate", "Mubarak Al-Kabeer Governorate", "al Ahmadi Governorate"),
stringsAsFactors = FALSE
)
geofacet::grid_preview(mygrid)

Bangladesh geofacet plot: Female population projections by upazilas

BGD_plot

Bangladesh 2020 population pyramids by upazila

BGD_pyr.png

Georgia Mapping in R

You can download session 9 files for constructing the population pyramids of Georgia here: https://github.com/rladies/meetup-presentations_tbilisi and specify your working directory with setwd(“/Users/mydomain/myfolder/”)

#set working directory
mypath<-"/Users/DrSpengler/The rectification of the Vuldrini/"
#upload shape files
georgia <- readOGR("./GEO_adm/","GEO_adm0")

## OGR data source with driver: ESRI Shapefile
## Source: "./GEO_adm/", layer: "GEO_adm0"
## with 1 features
## It has 70 fields

# plot(georgia, lwd=1.5)

georgia1 <- readOGR("./GEO_adm/","GEO_adm1")

## OGR data source with driver: ESRI Shapefile
## Source: "./GEO_adm/", layer: "GEO_adm1"
## with 12 features
## It has 16 fields

# plot(georgia1)

georgia2 <- readOGR("./GEO_adm/","GEO_adm2")

## OGR data source with driver: ESRI Shapefile
## Source: "./GEO_adm/", layer: "GEO_adm2"
## with 69 features
## It has 18 fields

# plot(georgia2)

gwat <- readOGR("./GEO_wat/" , "GEO_water_lines_dcw")

## OGR data source with driver: ESRI Shapefile
## Source: "./GEO_wat/", layer: "GEO_water_lines_dcw"
## with 559 features
## It has 5 fields

# plot(gwat)

gpop <- raster("./GEO_pop/geo_pop.grd")
# plot(gpop)

galt <- raster("./GEO_msk_alt/GEO_msk_alt.grd")
# plot(galt)

 plot(georgia, lwd=1.5) #n1

map1

 plot(georgia1, lwd=1.5) #n2

map2

 plot(georgia2, lwd=1.5) #n3

map3

 plot(georgia, lwd=1.5) #n4
 plot(gwat, lwd=1.5, col="blue", add=T) #n4

map4

 plot(gpop) #n5
 plot(georgia, lwd=1.5,  add=T) #n5

map5

 plot(galt, lwd=1.5) #n6

map6

Plot neighbouring countries

tur <- readOGR("./TUR_adm" , "TUR_adm0")

## OGR data source with driver: ESRI Shapefile
## Source: "./TUR_adm", layer: "TUR_adm0"
## with 1 features
## It has 70 fields
## Integer64 fields read as strings:  ID_0 OBJECTID_1

arm <- readOGR("./ARM_adm" , "ARM_adm0")

## OGR data source with driver: ESRI Shapefile
## Source: "./ARM_adm", layer: "ARM_adm0"
## with 1 features
## It has 70 fields
## Integer64 fields read as strings:  ID_0 OBJECTID_1

rus <- readOGR("./RUS_adm" , "RUS_adm0")

## OGR data source with driver: ESRI Shapefile
## Source: "./RUS_adm", layer: "RUS_adm0"
## with 1 features
## It has 70 fields
## Integer64 fields read as strings:  ID_0 OBJECTID_1

aze <- readOGR("./AZE_adm" , "AZE_adm0")

## OGR data source with driver: ESRI Shapefile
## Source: "./AZE_adm", layer: "AZE_adm0"
## with 1 features
## It has 70 fields
## Integer64 fields read as strings:  ID_0 OBJECTID_1

plot maps

plot(georgia, lwd=1.5, col="white", bg="lightblue")
plot(georgia1, add=T, lty=2)
plot(tur, add=T, col="white")
plot(arm, add=T, col="white")
plot(rus, add=T, col="white")
plot(aze, add=T, col="white")

map7

add labels for the countries

x.loc <- c(44.32002, 46.35746, 44.40421, 42.18156, 40.71662)
y.loc <- c(43.42472, 40.87209, 40.82228, 40.90945, 41.99276)
nb.lab <- c("Russia", "Azerbaijan", "Armenia", "Turkey", "Black Sea")
plot(georgia, lwd=1.5, col="white", bg="lightblue")
plot(georgia1, add=T, lty=2)
plot(tur, add=T, col="white")
plot(arm, add=T, col="white")
plot(rus, add=T, col="white")
plot(aze, add=T, col="white")
text(x.loc, y.loc, nb.lab)

let’s add everything (or almost everything) together

plot(gwat, col="blue")
# plot(georgia1[1,], lwd=1, col="lightblue", border="black", add=T)
plot(georgia2, lwd=0.5, border="black", lty=3, add=T)
plot(georgia1, border="black", lty=2, add=T)
plot(georgia, lwd=1.5, add=T)

map8

check georgia@data

head(georgia1)

##   ID_0 ISO  NAME_0 ID_1       NAME_1 VARNAME_1 NL_NAME_1 HASC_1 CC_1
## 0   81 GEO Georgia 1034     Abkhazia   Sokhumi      <NA>  GE.AB <NA>
## 1   81 GEO Georgia 1035       Ajaria    Batumi      <NA>  GE.AJ <NA>
## 2   81 GEO Georgia 1036        Guria  Ozurgeti      <NA>  GE.GU <NA>
## 3   81 GEO Georgia 1037      Imereti   Kutaisi      <NA>  GE.IM <NA>
## 4   81 GEO Georgia 1038      Kakheti    Telavi      <NA>  GE.KA <NA>
## 5   81 GEO Georgia 1039 Kvemo Kartli   Rustavi      <NA>  GE.KK <NA>
##                   TYPE_1           ENGTYPE_1 VALIDFR_1 VALIDTO_1 REMARKS_1
## 0 Avtonomiuri Respublika Autonomous Republic      1994   Present      <NA>
## 1 Avtonomiuri Respublika Autonomous Republic      1994   Present      <NA>
## 2                 Region              Region      1994   Present      <NA>
## 3                 Region              Region      1994   Present      <NA>
## 4                 Region              Region      1994   Present      <NA>
## 5                 Region              Region      1994   Present      <NA>
##   Shape_Leng Shape_Area
## 0   6.643211  0.9744622
## 1   3.055014  0.3074264
## 2   2.880653  0.2092665
## 3   4.214567  0.6783179
## 4   6.820519  1.2485036
## 5   5.219352  0.6807876

print labels on the map

labels for admin 2

coords2<- coordinates(georgia2[2:6,])
admin2 <- c(as.character(georgia2$NAME_2[1:5]))
admin2

## [1] "Gagra"      "Gali"       "Gudauta"    "Gulripshi"  "Ochamchire"

Upload data from World Bank

dt <- read.csv("/Users/ac1y15/Google Drive/blog/RLadies_Georgia_files/Session_3/Data_Extract_From_Subnational_Malnutrition/3f075abc-c51c-40c5-afb1-f8fbcfa30f23_Data.csv", header=T)
dt.1 <- subset(dt, dt$type==1&dt$select==1)

head(dt.1)

##            Admin.Region.Name select order
## 6                                 1     1
## 7  Georgia, Adjara Aut. Rep.      1     2
## 16            Georgia, Guria      1     3
## 26          Georgia, Imereti      1     4
## 31          Georgia, Kakheti      1     5
## 36     Georgia, Kvemo Kartli      1     6
##                         Admin.Region.Code type
## 6                                            1
## 7  GEO_Adjara_Aut._Rep._GE.AR_1297_GEO002    1
## 16            GEO_Guria_GE.GU_1298_GEO003    1
## 26          GEO_Imereti_GE.IM_1299_GEO004    1
## 31          GEO_Kakheti_GE.KA_1300_GEO005    1
## 36     GEO_Kvemo_Kartli_GE.KK_1301_GEO006    1
##                                                            Series.Name
## 6
## 7  Prevalence of overweight, weight for height (% of children under 5)
## 16 Prevalence of overweight, weight for height (% of children under 5)
## 26 Prevalence of overweight, weight for height (% of children under 5)
## 31 Prevalence of overweight, weight for height (% of children under 5)
## 36 Prevalence of overweight, weight for height (% of children under 5)
##          Series.Code YR2000 YR2005 YR2009
## 6                        NA     NA     NA
## 7  SN.SH.STA.OWGH.ZS     NA   28.1     NA
## 16 SN.SH.STA.OWGH.ZS     NA    7.9     NA
## 26 SN.SH.STA.OWGH.ZS    9.9   21.5     NA
## 31 SN.SH.STA.OWGH.ZS    7.0   19.6   13.2
## 36 SN.SH.STA.OWGH.ZS    9.5   28.2   19.1

Map the prevalence overweight w/h

library(classInt)
nclassint <- 3 #number of colors to be used in the palette
cat <- classIntervals(dt.1$YR2005, nclassint,style = "quantile") #style refers to how the breaks are created

colpal <- brewer.pal(nclassint,"Greens") #sequential
color.palette <- findColours(cat,colpal)
is.na(color.palette)

##  [1]  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE
## [12] FALSE

bins <- cat$brks
lb <- length(bins)

color.palette[c(1, 10)] <- "gray"
value.vec <- c(round(bins[-length(bins)],2))
value.vec.tail <- c(round(bins[-1],2))

Plot and SAVE map:

plot(georgia1, col=color.palette, border=T, main="Prevalence of overweight, \nweight for height (% of children under 5)")
legend("topright",fill=c("gray", "#E5F5E0", "#A1D99B", "#31A354"),legend=c("NA",paste(value.vec,":",value.vec.tail)),cex=1.1, bg="white", bty = "n")
# map.scale(41, 41, 2, "km", 2, 100)
map.scale(x=40.1, y=41.2, relwidth=0.1 , metric=T, ratio=F, cex=0.8)
SpatialPolygonsRescale(layout.north.arrow(2), offset= c(40.1, 41.6), scale = 0.5, plot.grid=F)

map12

See you at IUSSP to talk about the fantastic work we do at WorldPop! (plus Demotrends and R-Ladies)

As IUSSP is approaching, I’m looking forward to talk more about fine grid scale mapping research at WorldPop (University of Southampton) and Flowminder.

I will present my research on High resolution mapping of fertility and mortality from national household survey data in low income settings the 30th of October in session 5 at 8:30am Integrating spatial and statistical methods in demographic research, meeting room 1.41 and 1.42.

Prof. Andy Tatem will host a side meeting Geospatial Demography: Combining Satellite, Survey, Census and Cellphone Data to Provide Small-area Estimates on the 29 October 2017, 8:30-16:00.

and  will be contributing to the Cape Town R-Ladies chapter Saturday 4 November (details here) with a talk on R-Ladies and data visualization in ggplot2. Come talk to us and become an R-Lady, we are looking forward to sharing our experiences.

A few Demotrend(ers) will also be presenting at IUSSP, come talk to us 🙂

See you in Cape Town!