Dewey Dunnington 11/19/2016
Libraries
library(prettymapr)
library(dplyr)
library(ggplot2)
library(ggspatial)
library(reshape2)Load in data
mt <- read.csv('data/monthly_totals.csv', stringsAsFactors = FALSE)
mt$posix <- lubridate::dmy(mt$Date)
mt$posix[is.na(mt$posix)] <- lubridate::mdy(mt$Date[is.na(mt$posix)])
mt$month <- paste(lubridate::year(mt$posix),sprintf('%0.2d', lubridate::month(mt$posix)), sep="-")
mt$Items <- tolower(mt$Items)
mt$Supplied[mt$Supplied != "Y"] <- "N"
mt$Supplied <- factor(mt$Supplied, levels=c("Y", "N"))
mtgeo <- read.csv('data/monthly_totals_postals.csv')
deliveries <- mt %>% group_by(month, posix, Delivery.ID) %>%
summarise(Geo=unique(Geo), nitems=length(Items),
nyes=sum(Supplied=="Y"), nno=sum(Supplied=="N")) %>%
merge(mtgeo, by='Geo', all.x=T) %>%
select(posix, month, Delivery.ID, nitems, nyes, nno, postal, lon, lat)Plot geo
deliveriessum <- deliveries %>% group_by(postal, lon, lat) %>%
summarise(n=sum(nitems), fulfilled=sum(nyes)/sum(nitems))
ggplot(deliveriessum, aes(lon, lat)) + geom_osm() +
geom_spatial(mapping=aes(size=n, col=fulfilled)) + coord_fixed() +
scale_x_continuous(expand = c(0.05, 0.05)) +
scale_y_continuous(expand = c(0, 0.05))
Plot geo, facet by month
deliveriessum <- deliveries %>% group_by(postal, lon, lat, month) %>%
summarise(n=sum(nitems), fulfilled=sum(nyes)/sum(nitems))
ggplot(deliveriessum, aes(lon, lat)) + geom_osm() +
geom_spatial(mapping=aes(size=n, col=fulfilled)) +
facet_wrap(~month) + coord_fixed() +
scale_x_continuous(expand = c(0.05, 0.05)) +
scale_y_continuous(expand = c(0.05, 0.05))
Tabular summary by items
mtsum <- mt %>% group_by(Items) %>%
summarise(SuppliedY=sum(Supplied=="Y"),
SuppliedN=length(Supplied)-SuppliedY,
Total=length(Supplied))
mtsumtot <- rbind(mtsum[order(mtsum$Total, decreasing = T),], data.frame(Items="Totals", SuppliedY=sum(mtsum$SuppliedY),
SuppliedN=sum(mtsum$SuppliedN),
Total=sum(mtsum$Total)))
# knitr::kable(mtsumtot)Graphical summary by items
# plot by order of item frequency
mtsummelt <- mt %>% select(Items, Supplied, month)
mtsummelt$Items <- factor(mtsummelt$Items, levels=mtsum$Items[order(mtsum$Total)])
ggplot(mtsummelt, aes(x=Items, fill=Supplied)) + coord_flip() + stat_count()
# plot proportion
ggplot(mtsummelt, aes(x=Items, fill=Supplied)) + coord_flip() + stat_count(position='fill')
# plot by order of proportion
mtsummelt$Items <- factor(mtsummelt$Items, levels=mtsum$Items[order(mtsum$SuppliedY/mtsum$Total)])
ggplot(mtsummelt, aes(x=Items, fill=Supplied)) + coord_flip() + stat_count(position='fill')
mtsummelt$Items <- factor(mtsummelt$Items, levels=mtsum$Items[order(mtsum$Total)])
ggplot(mtsummelt, aes(x=Items, fill=Supplied)) + stat_count() + coord_flip() +
facet_wrap(~month) + theme_grey(8)
Graphical summary by date
# plot deliveries over time
deliverymelt <- deliveries %>% melt(id.vars=c("posix", "Delivery.ID", "nitems", "postal"),
measure.vars=c("nyes", "nno"), value.name="itemcount",
variable.name="Supplied")
ggplot(deliverymelt, aes(x=posix, y=itemcount, fill=Supplied)) + geom_bar(stat='identity')