Visualising UK Covid-19 Data - Part 3

R ggplot Data Analysis Data Visualisation Charting Covid-19

A recent news report suggested that Covid-19 cases in England and Wales are increasing once more. Here we look at data for 2023 and show the data on a UK map

Graham Cox
2023-04-11

Overview

In part two we downloaded data from the UK Government’s Health Security Agency Covid-19 Dashboard website.

Load Data

The data is now only updated weekly, so a CSV file with the data was saved to remove the need for daily downloads of data.

df <- read_csv("all_areas.csv")
# A tibble: 6 × 5
  area_code area_name        date       new_cases_by_specime…¹ country
  <chr>     <chr>            <date>                      <dbl> <chr>  
1 E06000003 Redcar and Clev… 2023-04-12                      6 England
2 E06000014 York             2023-04-12                      3 England
3 E06000050 Cheshire West a… 2023-04-12                      4 England
4 E08000001 Bolton           2023-04-12                     10 England
5 E08000016 Barnsley         2023-04-12                      4 England
6 E08000031 Wolverhampton    2023-04-12                      4 England
# ℹ abbreviated name: ¹​new_cases_by_specimen_date

Transform Data

Top Ten Areas

For the plots that will be created for this post, we need to look for the top five areas, or counties, in England and Wales, so we can plot the labels to the final map.

top_ten_all_areas_df <- df %>%
  
  # Filter data for just 2023
  filter(date >= as.Date("2023-01-01")) %>%
  
  # Group by area_name variable
  group_by(area_name) %>%
  
  # Calculate total number of cases by area_name
  summarise(cases = sum(new_cases_by_specimen_date)) %>%
  
  # Sort in descending order
  arrange(-cases) %>%
  
  # Get the top five items
  mutate(lbl = if_else(row_number() <= 5, TRUE, FALSE)) 
# A tibble: 203 × 3
   area_name       cases lbl  
   <chr>           <dbl> <lgl>
 1 Hampshire       10105 TRUE 
 2 Kent             8691 TRUE 
 3 Essex            8425 TRUE 
 4 Lancashire       8168 TRUE 
 5 Surrey           7109 TRUE 
 6 Hertfordshire    6754 FALSE
 7 Staffordshire    6221 FALSE
 8 Norfolk          5615 FALSE
 9 Nottinghamshire  5563 FALSE
10 Derbyshire       5423 FALSE
# ℹ 193 more rows

Create Plot

Create Plot Labels

We can now add some labels with some descriptive text explaining the chart to the viewer.

# Create title
title_text <- 
  "How many new cases of Covid-19 have been submitted?"

# Create subtitle
subtitle_text <- 
  paste(
    top_area_names,
    "are the top five counties with a total of",
    comma(top_area_counts),
    "cases, representing",
    top_area_pct,
    "of all cases in England & Wales since the start of 2023"
  )

# Create caption
caption_text <- "Source: UK Health Security Agency at
https://coronavirus.data.gov.uk/"

Create Colour Palette

Create a new palette of colours using the colorspace package to remove the standard ggplot2 colour palette.

# Create colour palette using the RedOr (Red-Orange) palette
pal <-
  colorspace::sequential_hcl(
    length(unique(df$area_name)), 
    palette = "redor", 
    rev = TRUE)

Join to Map Data

Now we have the data loaded, labels created and a new colour palette, we now need to join the data to the map data. Using the sf package, we can read the shape file saved in the same directory as the source data.

# Read the shape file into new data frame
counties <-
  sf::st_read(
    "Counties_and_Unitary_Authorities_(December_2016)_Boundaries.shp"
  )

Now we can join the data from the shape file to the data frame containing the data we want to map

counties <- counties %>% 
  inner_join(df, c("ctyua16cd" = "area_code"))

Create Plot

Basic plot

Now we have the data loaded, labels created and a new colour palette, let’s create our plot.

ggplot(counties) +
  geom_sf(aes(fill = cases), size = .4) +
  
  labs(title = title_text,
       subtitle = subtitle_text,
       caption = caption_text) +
  scale_fill_gradientn(
    colours = pal,
    labels = label_number(scale_cut = cut_short_scale()),
    guide = guide_colorbar(title = NULL)
  )
Top Ten Areas in England & Wales

Figure 1: Top Ten Areas in England & Wales

Amend theme and legend

Let’s remove the axis labels for the coordinates and amend the legend and move to the bottom of the plot.

ggplot(counties) +
  geom_sf(aes(fill = cases), size = .4) +
  
  labs(title = title_text,
       subtitle = subtitle_text,
       caption = caption_text) +
  scale_fill_gradientn(
    colours = pal,
    labels = label_number(scale_cut = cut_short_scale()),
    guide = guide_colorbar(title = NULL)
  ) +
  theme(
    plot.margin = margin(rep(.25, 4), unit = "cm"),
    plot.subtitle = element_textbox_simple(margin = margin(t = 1)),
    axis.text = element_blank(),
    axis.title = element_blank(),
    legend.position = "bottom",
    legend.direction = "horizontal",
    legend.key.height = unit(.8, units = "lines"),
    legend.key.width = unit(3, units = "lines")
  )
Top Ten Areas in England & Wales with amended legend

Figure 2: Top Ten Areas in England & Wales with amended legend

Add labels

For someone outside of the United Kingdom, the county names in the subtitle will be meaningless if the location of those counties are not known. Let’s add some labels for the top five counties using the ggrepel package.

ggplot(counties) +
  geom_sf(aes(fill = cases), size = .4) +
  
  labs(title = title_text,
       subtitle = subtitle_text,
       caption = caption_text) +
  scale_fill_gradientn(
    colours = pal,
    labels = label_number(scale_cut = cut_short_scale()),
    guide = guide_colorbar(title = NULL)
  ) +
  theme(
    plot.margin = margin(rep(.25, 4), unit = "cm"),
    plot.subtitle = element_textbox_simple(margin = margin(t = 1)),
    axis.text = element_blank(),
    axis.title = element_blank(),
    legend.position = "bottom",
    legend.direction = "horizontal",
    legend.key.height = unit(.8, units = "lines"),
    legend.key.width = unit(3, units = "lines")
  ) +
  ggrepel::geom_text_repel(
    data = counties %>% filter(lbl == TRUE),
    aes(x = long, y = lat, label = area_name),
    nudge_x = c(1.28, 0, 1.25, -1.5, .4),
    nudge_y = c(0,-.6, 0, 0, -.7),
    family = "roboto-condensed"
  )
Top Ten Areas in England & Wales with amended legend and county labels

Figure 3: Top Ten Areas in England & Wales with amended legend and county labels

Filtering for North East England

To filter the data for the area of the UK where I live, we can filter the data on the initial loading. First, create a vector of area names in North East England

# Create list of areas to use in the filter
areas <- c(
  "Hartlepool","Middlesbrough","Redcar and Cleveland",
  "Stockton-on-Tees","Darlington","County Durham",
  "Northumberland","Newcastle upon Tyne","North Tyneside",
  "South Tyneside","Sunderland","Gateshead"
  )

Apply the filter when the data is initially loaded.

The steps outlined in the previous sections can then be followed once more to create the plot below.

The values for the nudge_x parameter in the ggrepel::geom_text_repel need to be amended for the new top five counties. We can remove the nudge_y parameter as the areas to be labelled will be far enough apart without the need to vertically amend the location of the label.

nudge_x = c(0, 0, .6, .3, .5)
A facet plot of the Top Ten Areas in North East England

Figure 4: A facet plot of the Top Ten Areas in North East England

Conclusion

In this post we looked at the top ten areas in both England & Wales as a whole and the in the North East England and mapped the values to a map using the sf package.

The values used by the nudge_x and nudge_y parameters in the geom_text_repel were manually calculated by trial and error, amending the values with each iteration of preparing the images for this post. While not my preferred method, it was a good learning exercise.

In the next post, we will look at the actual numbers of cases submitted for Covid-19 as these values have not been shown in any of the previous posts.