IGFAE.IGFAECQD.R

#
options(stringsAsFactors=F)
options(digits=2)
#
#
# returns string w/o leading whitespace
trim.leading <- function (x)  sub("^\\s+", "", x)
#
# returns string w/o trailing whitespace		
trim.trailing <- function (x) sub("\\s+$", "", x)
#
# returns string w/o leading or trailing whitespace
trim <- function (x) gsub("^\\s+|\\s+$", "", x)
#
#
library(openxlsx) #to use read.xlsx()
library(readxl) #to read Excel's sheets names and *.xls files
library(dplyr) #to filter, order, rarrange...
#
##
## https://www.sharpsightlabs.com/blog/map-oil-production-country-r/
## as an example for creating maps
##
library(tidyverse)
library(sf)
library(rvest)
library(stringr)
library(scales)
library(maps)
#
map.world <- map_data('world')
#
#IGFAEQCD <- file.path("/Users/rrodriguez/Universidade de Santiago de Compostela/SALGADO LOPEZ CARLOS ALBERTO - IGFAE/GlobalTalent/IGFAEGlobalTalent2018/IGFAEQCD.appl.csv")
#IGFAEQCD <- file.path("C:/Users/Ricardo Rodríguez/Universidade de Santiago de Compostela/IGFAE-EB - Documents/0001.HumanCapital/Internal_postdoc/2018/MdM_QCD/appl-Original.xls")
IGFAEQCD <- file.path("C:/Users/ricar/Universidade de Santiago de Compostela/IGFAE-EB - Documents/HumanCapital/Scouting/MdM_QCD/appl-Original.xls")
#
IGFAEQCDdata <- readxl::read_excel(IGFAEQCD, sheet = "ajo154590681752867")
#
IGFAEQCDdata$LastUpdated <- as.Date(IGFAEQCDdata$LastUpdated)
IGFAEQCDdata$Received <- as.Date(IGFAEQCDdata$Received)
#
IGFAEQCDdata[,c('Uid','Name','Degree_Institution')]
#
as.data.frame(IGFAEQCDdata[,c('Uid')])
#
#AJO <- file.path("/Users/rrodriguez/Universidade de Santiago de Compostela/SALGADO LOPEZ CARLOS ALBERTO - IGFAE/GlobalTalent/IGFAEGlobalTalent2018/AJOuid.xlsx")
AJO <- file.path("C:/Users/Ricardo Rodríguez/Universidade de Santiago de Compostela/IGFAE-EB - Documents/0002.GlobalTalent/IGFAEGlobalTalent2018/AJOuid.xlsx")
#
AJOdata <- openxlsx::read.xlsx(AJO, sheet = "people")
#
IGFAEQCDdataAll <- merge(IGFAEQCDdata, AJOdata)
#
table(IGFAEQCDdataAll$gender)
#
prop.table(table(IGFAEQCDdataAll$gender))
#
anti_join(IGFAEQCDdataAll, map.world, by = c('countryDegree' = 'region'))
#
IGFAEQCDbyCountry <- as.data.frame(table(IGFAEQCDdataAll$countryDegree), stringsAsFactors = default.stringsAsFactors())
#
names(IGFAEQCDbyCountry) <- c('countryDegree','nApplications')
#
map.IGFAEQCD <- left_join(map.world, IGFAEQCDbyCountry, by = c('region' = 'countryDegree'))
#
anti_join(IGFAEQCDbyCountry, map.world, by = c('countryDegree' = 'region'))
#
ggplot(map.IGFAEQCD, aes( x = long, y = lat, group = group )) + geom_polygon(aes(fill = nApplications))
#