RStudio logo Leaflet logo Map emoji (Facebook) Thumbs-up emoji

1 Intro

1.1 Basic approach

We’re going to use R and Leaflet (a Javascript library) to plot schools on a very simple map and show the boundaries of local authority districts. We’re going to colour arbitrarily some of the LADs based on a particular condition and also alter the icon and colour of the markers according to Ofsted result, again arbitrarily.

We’re going to:

Read Local Authority District (LAD) polygon data (boundaries) and subset for England
Read Schools data (points)
Build a map with these components

Note that the code and the data is arbitrary and for training purposes only. This is intended purely as an introduction; there are probably better ways to do the things I’ve laid out below, but this should help give you the gist. See the Going Further and Resources sections for more detail on how to do these things and find some further examples.

1.2 Setup

I recommend setting up your project as an RStudio project for easy reproducibility and transferability. Ask if you want more information about this.

We’re going to make heavy use of the dplyr package for manipulation of our data and our maps. Remember when installing a package, you can use theinstall.packages() function to download it to your local library (a folder on your machine). You then only need to call the package from the library in future with the library() function.

library(dplyr)  # data manipulation and pipe operator (%>%)

Note also that I’m using the :: notation in the code below to show detail the package::function() relationship.

2 Local authority boundaries

2.1 Collect the data

Local authority district (LAD) polygon data is available from the ONS Open Geography Portal. We’re going to download a file of simplified LAD polygons from the URL using the Application Programming Interface (API) provided by ONS.

We’re downloading it as a GeoJSON file. This is a an open file type that is capable of holding geographical information and associated data, such as points, lines and polygons. The structure of a GeoJSON is nested and therefore differs from the ‘flat’ tables of a CSV file, for example.

# set the URL from where the file can be collected
geojson_url <- "https://opendata.arcgis.com/datasets/ae90afc385c04d869bc8cf8890bd1bcd_4.geojson"

# set the file path where we'll download to
file_path <- "data/regions.geojson"

# download from the URL to the filepath specified
download.file(geojson_url, file_path)

Now we can read the GeoJSON file into our environment.

lads <- geojsonio::geojson_read(
  x = file_path,  # the same file_path as above
  what = "sp"  # returns the read with spatial class
)

2.2 Manipulate the data

2.2.1 Spatial Polygons Data Frame

The GeoJSON is read as a special type of dataframe called a Spatial Polygons Data Frame (SPDF). An SPDF has special ‘slots’ for storing information. This includes a slot for co-ordinate data and a slot for the variables associated with those points. (Note that there’s also Spatial Points Data Frame that we’ll be dealing with later).

dplyr::glimpse(lads)

## Formal class 'SpatialPolygonsDataFrame' [package "sp"] with 5 slots
##   ..@ data       :'data.frame':  380 obs. of  10 variables:
##   .. ..$ objectid      : int [1:380] 1 2 3 4 5 6 7 8 9 10 ...
##   .. ..$ lad17cd       : Factor w/ 380 levels "E06000001","E06000002",..: 1 2 3 4 5 6 7 8 9 10 ...
##   .. ..$ lad17nm       : Factor w/ 380 levels "Aberdeen City",..: 148 198 247 309 86 139 346 25 26 171 ...
##   .. ..$ lad17nmw      : Factor w/ 23 levels " ","Abertawe",..: 1 1 1 1 1 1 1 1 1 1 ...
##   .. ..$ bng_e         : int [1:380] 447157 451141 464359 444937 428029 354246 362744 369490 332763 511894 ...
##   .. ..$ bng_n         : int [1:380] 531476 516887 519597 518183 515649 382146 388456 422806 436633 431716 ...
##   .. ..$ long          : num [1:380] -1.27 -1.21 -1.01 -1.31 -1.57 ...
##   .. ..$ lat           : num [1:380] 54.7 54.5 54.6 54.6 54.5 ...
##   .. ..$ st_areashape  : num [1:380] 9.66e+07 5.48e+07 2.47e+08 2.07e+08 1.98e+08 ...
##   .. ..$ st_lengthshape: num [1:380] 50252 35463 78677 86958 91352 ...
##   ..@ polygons   :List of 380
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. ..$ :Formal class 'Polygons' [package "sp"] with 5 slots
##   .. .. [list output truncated]
##   ..@ plotOrder  : int [1:380] 336 350 349 329 346 57 379 340 333 53 ...
##   ..@ bbox       : num [1:2, 1:2] -8.65 49.88 1.76 60.86
##   .. ..- attr(*, "dimnames")=List of 2
##   ..@ proj4string:Formal class 'CRS' [package "sp"] with 1 slot

The SPDF is nested so that the data (i.e. a data frame, a list, a vector) are stored in ‘slots’. You can see in the output above that the slots start with @ and then the items therein can (in general) be called with $ or [] notation. SO the structure of our SPDF looks like:

lads – our spatial polygons data frame
- @data – slot for variables associated with each of our polygons
  - $objectid – a variable with a unique object identification
  - $lad17cd – a variable with a unique local authority code
  - etc – other variables
- @polygons – slot for defining polygon shape for each local authority
  - $[first polygon] – the polygons themselves are nested
  - $[second polygon] – there’s as many polygons as there are LADS
  - etc
- @plotOrder – slot for the order in which the polygons are placed on the map
- @bbox – slot for the ‘bounding box’, i.e. the extent of the map
- @proj4string – slot for the projection system used

To further investigate this, you can click the magnifying glass in the environment pane, to the right of the SPDF name:

Environment pane

This will open a tab in the scripting pane:

Explore SPDF

You can click the blue arrows to show the levels of nesting. Note that the polygons in the @polygons slot are themselves nested quite deeply. A SPDF helps keep all this complex infomration in one object.

2.2.2 England only

We know which of the LADs are in England because their LAD code (ladcd17) starts with ‘E’ (‘W’ and ‘S’ for Wales and Scotland, surprise surprise). There are 326 LADs in England, so the resulting SPDF should be that long, having been reduced from 380 (the number of LADs in Brit). We can access information using @ to specify the slot and $ to specify the variable.

lads_eng <- subset(
  x = lads,  # our data
  subset = grepl(  # subset the data where the following pattern is matched
    x = lads@data$lad17cd,  # in this variable in this slot of this SPDF
    pattern = "^E"  # subset anything starting with 'E'
  )
)

length(lads_eng@data$lad17cd)  # check that the number of LADs is reduced

## [1] 326

Success. We now have only the LADs in England.

3 Schools

3.1 Collect the data

We start with the data for the schools that you want to map. At very least this would have to be a list of school Unique Reference Numbers (URNs) or Local Authority Establishment Numbers (LAESTABs). We can then join school infomration from Get Information About Schools (GIAS), including co-ordinates in the form of eastings and northings.

For this demo, I sampled randomly 500 primary schools and 500 secondaries from the GIAS dataset and selected a few columns of interest. I saved the resulting dataframe object as an RDS file, so we can use the readRDS() to read the data into our environment.

schools <- readRDS(file = "data/gias_sample.RDS")

Here’s what the dataframe structure looks like:

dplyr::glimpse(schools)

## Observations: 1,000
## Variables: 9
## $ urn           <chr> "131570", "116743", "109932", "143371", "106846"...
## $ laestab       <chr> "3833929", "8842155", "8692240", "9362880", "372...
## $ sch_name      <chr> "Shire Oak VC Primary School", "Walford Primary ...
## $ sch_type      <chr> "Voluntary controlled school", "Community school...
## $ phase         <chr> "Primary", "Primary", "Primary", "Primary", "Pri...
## $ pupil_count   <chr> "208", "196", "276", "633", "249", "93", "210", ...
## $ ofsted_rating <chr> "Good", "Good", "Good", "Requires improvement", ...
## $ easting       <int> 428142, 359307, 466035, 504170, 440147, 460189, ...
## $ northing      <int> 436290, 220990, 171981, 170873, 392804, 198702, ...

The column names are pretty self-explanatory. Note that the dataset is tidy – one school per row – and that we have unique identifiers (urn and laestab) and co-ordinates (easting and northing).

Sometimes co-ordinates can’t be matched to schools from GIAS. We can check for any schools that are missing eastings and northings using filter() and is.na(). In our case, we have one set of co-ordinates missing for a secondary school:

schools %>%
  dplyr::filter(is.na(easting) | is.na(northing))

## # A tibble: 1 x 9
##      urn laestab                           sch_name            sch_type
##    <chr>   <chr>                              <chr>               <chr>
## 1 145502 8034010 Winterbourne International Academy Academy sponsor led
## # ... with 5 more variables: phase <chr>, pupil_count <chr>,
## #   ofsted_rating <chr>, easting <int>, northing <int>

So let’s create a new data frame object with complete co-ordinates information and check we have one fewer secondary school:

schools_nona <- schools %>%
  dplyr::filter(!is.na(easting) | !is.na(northing))

table(schools_nona$phase)

## 
##   Primary Secondary 
##       500       499

3.2 Convert to latitude and longitude

There are many co-ordinate systems in use. The eastings and northings in GIAS are based on the system called the British National Grid (BNG). We’ll reproject our BNG data to latitude and longitude because it’s easier to handle in Leaflet and is the system used for our local authority district layer anyway.

To do this, I’m using some code developed by Alex Singleton, a Professor of Geographic Infomration Science at the University of Liverpool.

First we’ll load the relevant packages:

library(rgdal)  #  Geospatial Data Abstraction Library functions
library(geojsonio)  # deal with json file
library(sp)  # deal with spatial data

Let’s isolate the co-ordinate columns into a single dataframe object. I’m going to pre-emptively change the names of these columns to longitude and latitude, but note that they haven’t yet been converted; they’re still in the BNG projection.

schools_coords <- schools_nona %>%
  dplyr::transmute(  # create new columns and drop all the others
    easting = as.numeric(as.character(easting)), # make this text column numeric
    northing = as.numeric(as.character(northing))
  ) %>% 
  dplyr::rename(longitude = easting, latitude = northing)  # rename

print(schools_coords)

## # A tibble: 999 x 2
##    longitude latitude
##        <dbl>    <dbl>
##  1    428142   436290
##  2    359307   220990
##  3    466035   171981
##  4    504170   170873
##  5    440147   392804
##  6    460189   198702
##  7    532625   166192
##  8    395782   566384
##  9    515703   148686
## 10    413341   130471
## # ... with 989 more rows

And we’ll isolate the data columns as well.

schools_data <- schools_nona %>%
  dplyr::select(-easting, -northing)  # select all columns except the coords

print(schools_data)

## # A tibble: 999 x 7
##       urn laestab                               sch_name
##     <chr>   <chr>                                  <chr>
##  1 131570 3833929            Shire Oak VC Primary School
##  2 116743 8842155                 Walford Primary School
##  3 109932 8692240                   Calcot Junior School
##  4 143371 9362880             Riverbridge Primary School
##  5 106846 3722023             Meadow View Primary School
##  6 123030 9312455             Stadhampton Primary School
##  7 142246 3063406       St Mary's Catholic Infant School
##  8 122222 9292227                 Beaufront First School
##  9 125151 9363044         Powell Corderoy Primary School
## 10 126265 8652191 Salisbury, Manor Fields Primary School
## # ... with 989 more rows, and 4 more variables: sch_type <chr>,
## #   phase <chr>, pupil_count <chr>, ofsted_rating <chr>

And now we’ll merge these together into a Spatial Points Data Frane (we saw the polygon version of this when we read in the GeoJSON file of LADs earlier in this document).

schools_spdf <- sp::SpatialPointsDataFrame(  # create a SPDF
  coords = schools_coords,  # the school co-ordinates
  data = schools_data,  # the school data
  proj4string = CRS("+init=epsg:27700")  # BNG projection system
) %>% 
  sp::spTransform(CRS("+init=epsg:4326")) # reproject to latlong system

Here’s what the structure of our SPDF looks like:

dplyr::glimpse(schools_spdf)

## Formal class 'SpatialPointsDataFrame' [package "sp"] with 5 slots
##   ..@ data       :Classes 'tbl_df', 'tbl' and 'data.frame':  999 obs. of  7 variables:
##   .. ..$ urn          : chr [1:999] "131570" "116743" "109932" "143371" ...
##   .. ..$ laestab      : chr [1:999] "3833929" "8842155" "8692240" "9362880" ...
##   .. ..$ sch_name     : chr [1:999] "Shire Oak VC Primary School" "Walford Primary School" "Calcot Junior School" "Riverbridge Primary School" ...
##   .. ..$ sch_type     : chr [1:999] "Voluntary controlled school" "Community school" "Community school" "Academy converter" ...
##   .. ..$ phase        : chr [1:999] "Primary" "Primary" "Primary" "Primary" ...
##   .. ..$ pupil_count  : chr [1:999] "208" "196" "276" "633" ...
##   .. ..$ ofsted_rating: chr [1:999] "Good" "Good" "Good" "Requires improvement" ...
##   ..@ coords.nrs : num(0) 
##   ..@ coords     : num [1:999, 1:2] -1.574 -2.593 -1.051 -0.503 -1.397 ...
##   .. ..- attr(*, "dimnames")=List of 2
##   ..@ bbox       : num [1:2, 1:2] -5.27 50.23 1.71 55.76
##   .. ..- attr(*, "dimnames")=List of 2
##   ..@ proj4string:Formal class 'CRS' [package "sp"] with 1 slot

As before, note the slots are preceded by an @ symbol. If you want to access the phase column for example, you would need to type schools_spdf@data$phase. Compare this to schools$phase, which you’d use to access the phase data in the original data frame.

4 Create maps

Now we can bring the LAD polygons and the schools data together into a simple map using the leaflet package. This package wraps up some functions from the Leaflet JavaScript library.

Much of this material below is derived from the Leaflet for R page from the RStudio team.

First we’ll call the package from our library.

library(leaflet)  # leaflet mapping functions

Leaflet maps are built up in layers. Our approach is going to be to do the following:

Add a basemap (the underlying map that contains geographic features)
Add our polygons (boundaries of LADs)
Add our markers (points where our sampled schools exist)

Here is an artist’s rendering of layers, with the map at back, then boundaries (polygons) overlaid and finally the points:

Layers

Using pipe (%>%) notation, we can build these layers up one-by-one. This is much like building a ggplot object by building up layers using the +.

4.1 Underlying map

Generating the underlying map is easy. We can choose a variety of background maps from a number of providers, but Open Street Map is a good place to start.

map <- leaflet::leaflet() %>%
  leaflet::addProviderTiles(providers$OpenStreetMap)

map  # show the map

Now we’ll add a layer containing the LAD polygons.

4.2 LAD polygons

To show the boundaries, we could simply render the edges and leave the area of the polygon clear. We could also pick out individual LADs based on some variable. Arbitrarily, let’s fill any that are over a certain area, given by the st_areashape variable in the GeoJSON file. We can do this using an ifelse() statement for the fillOpacity argument in the addPolygons() function.

map_lad <- map %>%
  leaflet::addPolygons(
    data = lads_eng,  # LAD polygon data from geojson
    weight = 1,  # line thickness
    opacity = 1,  # line transparency
    color = "black",  # line colour
    fillOpacity = ifelse(  # conditional fill opacity
      test = lads_eng@data$st_areashape > 1E+09,  # if area is over this value
      yes = 0.5,  # then make it half-opaque
      no = 0  # otherwise make it entirely transparent
    ),
    fillColor = "red",
    label = ~lad17nm  # LAD name as a hover label
  )

map_lad  # show the map

Hover over the polygons in the map above to see the name of the LAD.

4.3 Marker points

So now we’ll add some markers that show where are schools are. Rather than mark all the points with exactly the same marker, we’re going to colour the markers according got Ofsted rating and give them icons that relate to their educaiton phase.

What levels of Ofsted rating do we have?

table(schools_spdf@data$ofsted_rating)

## 
##                 Good           Inadequate          Outstanding 
##                  607                   20                  193 
## Requires improvement   Serious Weaknesses     Special Measures 
##                  113                    1                    1

So let’s colour differently three separate groups:

‘Outstanding’
‘Good’
everything elese (‘Inadequate’, ‘Requires improvement’, ‘Serious Weaknesses’ and ‘Special Measures’).

We’re going to use the addAwesomeMarkers() function rather than the addMarkers() function. This is because it’s more… awesome. It allows you to tweak a lot of settings with your markers. For example, you can use the awesomeIcons() function in the icon argument to select from a range of different icons. By specifying the library, you can name an icon and it will automatically be used. In this example, we’re using the ‘ion’ library and the icons named ’ion-arrow-*-b’. You can preview the icons available in the ‘ion’ library by visitng the ionicons website. You need only to replace the icon argument in the awesomeIcons() function with the name of any of the icons shown on the website (click the icon for its name).

map_groups <- map_lad %>% 
  leaflet::addAwesomeMarkers(
    data = schools_spdf,  # our spatial points data frame
    icon = awesomeIcons(
      library = "ion",
      icon = ifelse(
        test = schools_spdf@data$ofsted_rating == "Primary",
        yes = "ion-arrow-down-b",  # down arrow for primary
        no = "ion-arrow-up-b"  # up arrow for secondary
      ),
      iconColor = "white",  # the icon's colour
      markerColor = ifelse(
        test = schools_spdf@data$ofsted_rating == "Outstanding",  # if this...
        yes = "blue",  # ...then colour blue
        no = ifelse(
          test = schools_spdf@data$ofsted_rating == "Good",  # if this...
          yes = "lightblue",  # ...then colour lightblue
          no = "red"  # colour remaining schools (Ofsted rating RI, I, SW or SM)
            )
          )
        )
      )
    
map_groups  # show the map

This is better than plain markers that are the same for all schools, but it’s difficult to see what’s going on. We can instead introduce the concept of ‘groups’ that can be toggled on or off.

4.4 Toggle groups

4.4.1 Subset data

We’re going to do something similar to the map above, but each group of markers will be added separately with its own call to the addAwesomeMarkers() function.

First we need to subset the data into our categories of interest.

# Outstanding
ofsted_outst <- subset(
  x = schools_spdf,
  subset = schools_spdf@data$ofsted_rating == "Outstanding"
)

# Good
ofsted_good <- subset(
  x = schools_spdf,
  subset = schools_spdf@data$ofsted_rating == "Good"
)

# The others
ofsted_other <- subset(
  x = schools_spdf,
  subset = schools_spdf@data$ofsted_rating %in% c(
    "Requires Improvement", "Inadequate",
    "Serious Weaknesses", "Special Measures"
  )
)

4.4.2 Function: popups

While we’re here, we’re also going to add a function that will populate the on-click marker popups with detail about that individual school. Words in quotes will be rendered verbatim and personalised content will fill the popup where we’ve specified data from the @data slot of our SPDF. Note that we can use HTML in here too: <br> inserts a line break.

get_popup_content <- function(ofsted_group_spdf) {
  paste0(
    "<b>", ofsted_group_spdf@data$sch_name, "</b>",
    "<br>LAESTAB: ", ofsted_group_spdf@data$laestab,
    "<br>",
    "<br>Phase: ", ofsted_group_spdf@data$phase,
    "<br>Ofsted: ", ofsted_group_spdf@data$ofsted_rating,
    "<br>Pupils: ", ofsted_group_spdf@data$pupil_count, " (2017)"
  )
}

4.4.3 Function: add markers

And now we’ll create a function for adding the markers. Note that the get_popup_content() function we defined above is used here in the popup argument to addAwesomeMarkers()

assign_markers <- function(
  map_object,
  ofsted_group_spdf,
  group_name,
  marker_col = "blue"  # will default to blue
) {
  
  leaflet::addAwesomeMarkers(
    map = map_object,  # the base map we created first
    data = ofsted_group_spdf,  # the spatial points data frame
    group =  group_name,  # the name for the marker grouping
    popup = ~get_popup_content(ofsted_group_spdf),  # our popup function
    icon = awesomeIcons(
      library = "ion",  # the ion set of icons
      icon = ifelse(  # conditional icon
        test = ofsted_group_spdf@data$phase == "Primary",
        yes = "ion-arrow-down-b",  # primary gets a down arrow
        no = "ion-arrow-up-b"  # up arrows for secondary schools
      ),
      iconColor = "white",
      markerColor = marker_col  # you can specify a colout for the marker 
    )
  )  
}

4.4.4 Map

Now put it together. See how using out assign_marker() function – with the get_popup_content() function inside – helps cut down the code and make it more readable.

The ‘Oustatnding’ group are turned on by default and the other two are off. You can switch them on using the checkboxes in the top right. Note that the more markers that are showing, the slower it is to respond to your interactions.

map_toggle <- map_lad %>% 
  # add the markers (order is the order the checkboxes will appear on the map)
  assign_markers(  # marker group 1: 'outstanding' schools
    ofsted_group_spdf = ofsted_outst,  # the subset we specified earlier
    group_name = "Outstanding"  # sensible group name
  ) %>% 
  assign_markers(  # marker group 2: 'good' schools
    ofsted_group_spdf = ofsted_good,   # the subset we specified earlier
    group_name = "Good",  # sensible group name
    marker_col = "lightblue"
  ) %>% 
  assign_markers(  # marker group 3: schools performing worse than 'good'
    ofsted_group_spdf = ofsted_other,  # the subset we specified earlier
    group_name = "Other",  # sensible group name
    marker_col = "red"
  ) %>% 
  # controlling the groups
  leaflet::addLayersControl(
    overlayGroups = c("Outstanding", "Good", "Other"),  # add these layers
    options = layersControlOptions(collapsed = FALSE)  # expand on hover?
  ) %>% 
  hideGroup(c("Good", "Other"))  # turn these off by default
    
map_toggle  # show the map

The ‘good’ layer is still a bit busy, but we can remedy that.

4.4.5 Clusters

We can automatically group nearby markers into clusters, shown as coloured circles on the map. The number on the circle indicates the number of markers grouped together into that circle’ (see image below); higher numbers make the circle red, lower numbers make it green. Hovering over the circle shows you the extent of the markers demarcated with a displaying a blue boundary. As you zoom in, the clusters expand into smaller clusters and ultimately into individual markers.

Cluster of 12 schools

I’ve turned on the ‘Good’ group by default to make the clusters obvious. I’ve also built the points layer directly on top of the base map without the local authority boundaries, again for ease of understanding.

map_clusters <- map %>% 
  # add the markers (order is the order the checkboxes will appear on the map)
  assign_markers(  # use our function for adding markers
    ofsted_group_spdf = ofsted_outst,
    group_name = "Outstanding"
  ) %>%
  leaflet::addAwesomeMarkers(  # using addAwesomeMarkers function so you can...
    data = ofsted_good,
    clusterOptions = markerClusterOptions(),  # ...see how clusters are added
    group =  "Good",
    popup = ~get_popup_content(ofsted_good),
    icon = awesomeIcons(
      library = "ion",
      icon = ifelse(
        test = ofsted_good@data$phase == "Primary",
        yes = "ion-arrow-down-b",
        no = "ion-arrow-up-b"
      ),
      iconColor = "white",
      markerColor = "lightblue"
    )
  ) %>%
  assign_markers(  # use our function for adding markers
    ofsted_group_spdf = ofsted_other,
    group_name = "Other",
    marker_col = "red"
  ) %>% 
  # controlling the groups
  leaflet::addLayersControl(
    overlayGroups = c("Outstanding", "Good", "Other"),  # add these layers
    options = layersControlOptions(collapsed = FALSE)  # expand on hover?
  ) %>% 
  hideGroup(c("Outstanding", "Other"))  # turn these off by default
    
map_clusters # show the map

5 Going further

You can:

embed this in an R Shiny interface and have reactive elements (e.g. zoom to a LAD)
embed these maps in an HTML for sharing (like this one)
make a choropleth map by colouring your polygons on a continuous scale
change the marker style, not just the icon
plot shapes, e.g. circles of fixed radius from a point of origin

And much more.

Basic `leaflet` maps in R

EYSSAR divisional analysts meeting

Matt Dray

13 February 2018

1 Intro

1.1 Basic approach

1.2 Setup

2 Local authority boundaries

2.1 Collect the data

2.2 Manipulate the data

2.2.1 Spatial Polygons Data Frame

2.2.2 England only

3 Schools

3.1 Collect the data

3.2 Convert to latitude and longitude

4 Create maps

4.1 Underlying map

4.2 LAD polygons

4.3 Marker points

4.4 Toggle groups

4.4.1 Subset data

4.4.2 Function: popups

4.4.3 Function: add markers

4.4.4 Map

4.4.5 Clusters

5 Going further

6 Resources

Basic leaflet maps in R

EYSSAR divisional analysts meeting

Matt Dray

13 February 2018

1 Intro

1.1 Basic approach

1.2 Setup

2 Local authority boundaries

2.1 Collect the data

2.2 Manipulate the data

2.2.1 Spatial Polygons Data Frame

2.2.2 England only

3 Schools

3.1 Collect the data

3.2 Convert to latitude and longitude

4 Create maps

4.1 Underlying map

4.2 LAD polygons

4.3 Marker points

4.4 Toggle groups

4.4.1 Subset data

4.4.2 Function: popups

4.4.3 Function: add markers

4.4.4 Map

4.4.5 Clusters

5 Going further

6 Resources

Basic `leaflet` maps in R