When related data are stored in distinct data frames, it is possible to merge them into a single data frame.

The ..._join() methods take two data frames and produces a new one

• including all columns from the two data frames
• common (merged by) columns appear only once
• rows from the data frames are matched by
  • common columns
  • columns specified with parameter by

• inner_join() : includes only matching rows
• left_join() : includes all rows from left df + matching rows
• right_join() : includes all rows from right df + matching rows
• full_join() : includes all rows from both dfs

Two data frames:

• df1: id and name
• df2: id and day

df1 %>% inner_join(df2, by="id") %>% knitr::kable()

• Rows from dataframes are matched by id
• Rows in either data frame with no corresponding id in the other are discarded.

df1 %>% left_join(df2, by="id") %>% knitr::kable()

• Rows from dataframes are matched by id
• All rows in left data frame are included
• Rows in right data frame with no corresponding id in the left are discarded.

df1 %>% right_join(df2, by="id") %>% knitr::kable()

• Rows from dataframes are matched by id
• All rows in right data frame are included
• Rows in left data frame with no corresponding id in the right are discarded.

df1 %>% full_join(df2, by="id") %>% knitr::kable()

• Rows from dataframes are matched by id
• All rows from both data frames are included

There are several packages in R that allow drawing maps:

• ggplot2 using geom_sf()
• mapview interactive web-oriented maps
• leaflet based on the leaflet Javascript libray

Often when showing maps we deal with shapes representing geografic regions.

Geografic shapes are often shared using the Shapefile file format.

• Usually consist of a main .shp file plus .dbf, .prj, .shx files

Often available online for official boundaries

• Italian administrative boundaries published by ISTAT
  • https://www.istat.it/it/archivio/222527
• Worldwide free spatial data at DIVA-GIS
  • https://diva-gis.org/Data

Vector data is often encoded (internally) using the “simple features” standard

In R it is a dataframe containing a column named geometry

• Library sf can be used to read and manipulate
• geom_sf() in ggplot2 draws the layer
• coord_sf() predefined geographical coordinate system

it <- read_sf("Reg01012021_g/Reg01012021_g_WGS84.shp")
knitr::kable(head(it[,3:6],4))

ggplot(it,aes(geometry=geometry))+geom_sf()

ggplot(it,aes(fill=DEN_REG))+geom_sf()

Source: http://dati.istat.it/Index.aspx?QueryId=18460#

it %>% inner_join(pop_it,by=c("DEN_REG"="Territorio")) %>% 
ggplot(aes(fill=(maschi+femmine)/1000000))+geom_sf()+
  scale_fill_distiller(name="Population\n(milions)")

Plots can be combined using the library patchwork

It works by combining ggplot2 objects with operators:

• g1 + g2 : places the plots side by side
• g1 / g2 : places the plots one over the other
• ( and ): groups plots

md <- it %>% inner_join(pop_it,by=c("DEN_REG"="Territorio"))
pfm <- ggplot(md, aes(fill=maschi+femmine))+geom_sf()+
  scale_fill_distiller(palette="Greys",
                       labels=scales::label_number())
pf <- ggplot(md, aes(fill=femmine))+geom_sf()+
  scale_fill_distiller(palette="Reds",
                       labels=scales::label_number())
pm <- ggplot(md, aes(fill=maschi))+geom_sf()+
  scale_fill_distiller(palette="Blues",
                       labels=scales::label_number())
pfm + ( pf / pm )

• Nico Hahn, Making Maps with R
  • https://bookdown.org/nicohahn/making_maps_with_r5/docs/introduction.html
• Hadley Wickham, Danielle Navarro, and Thomas Lin Pedersen. “ggplot2: Elegant Graphics for Data Analysis”, in-progress
  • https://ggplot2-book.org/