The data we will be using throughout this course relates to the Core Spending Power (CSP) of English local authorities between 2015 and 2020. This is a measure of the resources available to local authorities in England to fund service delivery. The CSP is broken down into several components, presented as variables in the data. These components include:
+
+
Settlement Funding Assessment (sfa)
+
Compensation for under-indexing the business rates multipler (under_index)
+
Income from council tax (ct_total)
+
New Homes Bonus (nhb)
+
Rural Services Delivery Grant (rsdg)
+
+
Spending power is given in millions of pounds (£). The data were provided by the UK government’s Department for Levelling Up, Housing and Communities. Full guidance on the data can be found on the Department’s website. A brief description of the variables included in the data are given below.
+
+
+
Descriptions of variables
+
+
Identifier variables
+
Each dataset contains a unique identifier code variable, ons_code. This is a code given by the Government’s Office for National Statistics (ONS), and is used to join different datasets. There is also an authority variable which contains the local authority name (to see where each local authority lies on a map, you can visit the Government’s geoportal website).
+
+
+
Regions of England
+
In addition to each local authority’s unique code and name, we are given the region that they lie within. England is separated into 9 regions (shown on this map) which are given as acronyms in the data. These are:
+
+
L = London
+
NW = North West
+
NE = North East
+
YH = Yorkshire and the Humber
+
WM = West Midlands
+
EM = East Midlands
+
EE = East England
+
SW = South West
+
SE = South East
+
+
+
+
Settlement Funding Assessment (SFA)
+
The Settlement Funding Assessment (sfa in the data) is the baseline funding level of local authorities, and includes the Revenue Support Grant (a central government grant given to local authorities).
+
+
+
Under-indexing business rate multipliers
+
The under_index variable is given to compensate local authorities that under-indexed business rate multipliers in previous years (i.e. those that used a measure of inflation that was lower than that should have been used).
+
+
+
Council tax
+
Council tax (ct_total) is the income made by each local authority from council tax. In England, the amount of council tax charged to residents is set by each local authority to make up additional revenue needed to cover planned spending.
+
+
+
New Homes Bonus
+
The nhb variables is the funding received as part of the New Homes Bonus, a government inncentive to encourage local authorities to promote new housing delevopment.
+
+
+
Rural Services Delivery Grant
+
The rsdg variable is funding received as part of the Rural Services Delivery Grant, provided to rural councils to recognise additional costs in these areas.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/_book/data_description.qmd b/_book/data_description.qmd
deleted file mode 100644
index e712980..0000000
--- a/_book/data_description.qmd
+++ /dev/null
@@ -1,45 +0,0 @@
-# Data description {.unnumbered}
-
-## What is 'CSP'?
-The data we will be using throughout this course relates to the Core Spending Power (CSP) of English local authorities between 2015 and 2020. This is a measure of the resources available to local authorities in England to fund service delivery. The CSP is broken down into several components, presented as variables in the data. These components include:
-
-- Settlement Funding Assessment (`sfa`)
-- Compensation for under-indexing the business rates multipler (`under_index`)
-- Income from council tax (`ct_total`)
-- New Homes Bonus (`nhb`)
-- Rural Services Delivery Grant (`rsdg`)
-
-Spending power is given in millions of pounds (£). The data were provided by the UK government's Department for Levelling Up, Housing and Communities. Full guidance on the data can be found on the [Department's website](https://www.gov.uk/government/publications/explanatory-note-on-core-spending-power-provisional-local-government-finance-settlement-2024-to-2025/explanatory-note-on-core-spending-power-provisional-local-government-finance-settlement-2024-to-2025). A brief description of the variables included in the data are given below.
-
-## Descriptions of variables
-
-### Identifier variables
-Each dataset contains a unique identifier code variable, `ons_code`. This is a code given by the Government's Office for National Statistics (ONS), and is used to join different datasets. There is also an `authority` variable which contains the local authority name (to see where each local authority lies on a map, you can visit the [Government's geoportal website](https://geoportal.statistics.gov.uk/documents/cb64eeb1b0a74e5ca277f9fac58500f4/explore)).
-
-### Regions of England
-In addition to each local authority's unique code and name, we are given the region that they lie within. England is separated into 9 regions (shown on [this map](https://geoportal.statistics.gov.uk/documents/bfe6f7099f9e4cc59f8064962959bfb5/explore)) which are given as acronyms in the data. These are:
-
-- `L` = London
-- `NW` = North West
-- `NE` = North East
-- `YH` = Yorkshire and the Humber
-- `WM` = West Midlands
-- `EM` = East Midlands
-- `EE` = East England
-- `SW` = South West
-- `SE` = South East
-
-### Settlement Funding Assessment (SFA)
-The Settlement Funding Assessment (`sfa` in the data) is the baseline funding level of local authorities, and includes the Revenue Support Grant (a central government grant given to local authorities).
-
-### Under-indexing business rate multipliers
-The `under_index` variable is given to compensate local authorities that under-indexed business rate multipliers in previous years (i.e. those that used a measure of inflation that was lower than that should have been used).
-
-### Council tax
-Council tax (`ct_total`) is the income made by each local authority from council tax. In England, the amount of council tax charged to residents is set by each local authority to make up additional revenue needed to cover planned spending.
-
-### New Homes Bonus
-The `nhb` variables is the funding received as part of the New Homes Bonus, a government inncentive to encourage local authorities to promote new housing delevopment.
-
-### Rural Services Delivery Grant
-The `rsdg` variable is funding received as part of the Rural Services Delivery Grant, provided to rural councils to recognise additional costs in these areas.
diff --git a/_book/exercise_solutions.html b/_book/exercise_solutions.html
new file mode 100644
index 0000000..2d5bf4d
--- /dev/null
+++ b/_book/exercise_solutions.html
@@ -0,0 +1,1165 @@
+
+
+
+
+
+
+
+
+
+Introduction to R with Tidyverse - Exercise solutions
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Exercise solutions
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Exercise 1
+
+
Open a new script file if you have not already done so.
+
Save this script file into an appropriate location.
+
+
+
+
Solutions
+
+
To open a new R script, click the icon and select ‘R script’.
+
Save this file by following File -> Save as… from the drop-down menu, selecting the folder you are working from in this course, and giving the file an appropriate name.
+
+
+
+
Exercise 2
+
+
Add your name and the date to the top of your script file (hint: comment this out so R does not try to run it)
+
Use R to calculate the following calculations. Add the result to the same line of the script file in a way that ensures there are no errors in the code.
+
+
+
\(64^2\)
+
\(3432 \div 8\)
+
\(96 \times 72\)
+
+
When you have finished this exercise, select the entire script file (using Ctrl + a on windows or Command + a on Mac) and run it to ensure there are no errors in the code.
+
+
+
Solutions
+
+
Add a # symbol to the script file before printing your name and the date,
+
After running the calculation, copy and paste the result after a # symbol to ensure there are no errors:
How many local authorities were included in the London region?
+
Give three different ways that it would be possible to select all spend variables (sfa_2020, nhb_2020, etc.) from the CSP_2020 dataset.
+
Create a new tibble, em_2020, that just includes local authorities from the East Midlands (EM) region.
+
+
+
How many local authorities in the East Midlands had an SFA of between £5 and 10 million?
+
Create a new variable with the total overall spend in 2020 for local authorities in the East Midlands.
+
+
+
+
Solutions
+
+
First filter the data to return only rows which belong to the London region, then count the number of rows in this subgroup:
+
+
+
csp_2020 %>%
+filter(region =="L") %>%
+count()
+## # A tibble: 1 × 1
+## n
+## <int>
+## 1 34
+
+
+
There are many different ways to select variables from a dataset. For a list of selection helpers, check the helpfile ?tidyr_tidy_select. Some example include:
+
+
+
# Using the : symbol to return consecutive columns
+
+# By variable name:
+select(csp_2020, sfa_2020:rsdg_2020)
+
+# Or by column number:
+select(csp_2020, 4:9)
+
+# Returning all variables with names ending "_2020"
+select(csp_2020, ends_with("_2020"))
+
+# Return all numeric variables
+select(csp_2020, where(is.numeric))
+
+# Return all variables that are not character
+select(csp_2020, where(!is.character))
+
+
+
To create a new tibble, use filter to select the subgroup where region is “EM”, and attach as an object using the <- symbol
+
+
+
em_2020 <-filter(csp_2020, region =="EM")
+
+
+
Use filter to select the subgroup and then count the number of rows:
Use the mutate function to create a new variable from the existing ones. Hint: If you are not sure of the variable names in the data, use the names function and copy them from the console:
Create a data frame with the minimum, maximum and median total spend per year for each region.
+
Produce a frequency table containing the number and percentage of local authorities in each region.
+
Convert the data object csp_long2 back into wide format, with one row per local authority and one variable per total spend per year (HINT: start by selecting only the variables you need from the long data frame). Use the help file ?pivot_wider and vignette("pivot") for more hints.
+
Using your new wide data frame, calculate the difference in total spending for each local authority between 2015 and 2020. How many local authorities have had an overall reduction in spending since 2015?
+
+
+
+
Solutions
+
+
Use the summarise function after grouping by the year and region variables:
+
+
+
csp_long2 %>%
+group_by(region, year) %>%
+summarise(min_spend =min(total_spend),
+max_spend =max(total_spend),
+median_spend =median(total_spend)) %>%
+ungroup()
+## # A tibble: 54 × 5
+## region year min_spend max_spend median_spend
+## <chr> <dbl> <dbl> <dbl> <dbl>
+## 1 EE 2015 0 883. 15.9
+## 2 EE 2016 0 860. 16.2
+## 3 EE 2017 0 845. 15.0
+## 4 EE 2018 0 860. 14.4
+## 5 EE 2019 0 874. 14.7
+## 6 EE 2020 0 915. 15.2
+## 7 EM 2015 0 492. 12.4
+## 8 EM 2016 0 479. 12.0
+## 9 EM 2017 0 475. 11.1
+## 10 EM 2018 0 483. 11.0
+## # ℹ 44 more rows
+
+
+
To calculate the percentage of local authorities in each region, we need the total number in each region and the overall number of local authorities:
+
+
+
# Use the csp_2020 data as only require one row per local authority
+csp_2020 %>%
+# Begin by calculating number of local authorities per region
+group_by(region) %>%
+# Count number of rows in each group
+summarise(n_la_region =n()) %>%
+ungroup() %>%
+# Create a new variable with the total number of local authorities (the sum)
+mutate(n_la_overall =sum(n_la_region),
+# Calculate the percentage (and round to make easier to read)
+perc_la_region =round((n_la_region / n_la_overall) *100, 2)) %>%
+# Remove the total local authority column
+select(-n_la_overall)
+## # A tibble: 9 × 3
+## region n_la_region perc_la_region
+## <chr> <int> <dbl>
+## 1 EE 57 14.4
+## 2 EM 51 12.9
+## 3 L 34 8.59
+## 4 NE 15 3.79
+## 5 NW 46 11.6
+## 6 SE 81 20.4
+## 7 SW 47 11.9
+## 8 WM 38 9.6
+## 9 YH 27 6.82
+
+
+
Use the pivot_wider function, use the year to set the new variable names suffix (names_from =), add a prefix to avoid variable names beginning with a number (names_prefix =), and take the values_from the current total_spend column:
Begin by using mutate to create a variable with the difference between total spend 2015 - 2020. Use filter to return rows where there is a reduction in spend, count the number of rows:
Create a new data object containing the 2020 CSP data without the Greater London Authority observation. Name this data frame csp_nolon_2020.
+
Using the csp_nolon_2020 data, create a data visualisation to check the distribution (or shape) of the SFA variable.
+
Based on the visualisation above, create a summary table for the SFA variable containing the minimum and maximum, and appropriate measures of the centre/average and spread.
+
+
+
+
Solutions
+
+
Create a new object using the <- symbol, use filter to remove the duplicate row:
+
+
+
csp_nolon_2020 <-filter(csp_2020, authority !="Greater London Authority")
+
+
+
Histograms are used to check the distribution of numeric variables:
Use an appropriate data visualisation to show how the total spend in each local authority has changed over the years between 2015 and 2020. Choose a visualisation that shows these trends over time and allows us to compare them between regions.
+
+
+
Solution
+
The most appropriate plot to show a change in variable over time is a line graph (with year on the x-axis and spend on the y-axis). To compare these between regions, we could set the colour of these lines, but as there are so many local authorities, this would overload the graph and make it hard to compare. As an alternative, we can facet this graph by region to show the line graphs on the same scale on the same output.
+
Be sure to set appropriate axis labels, font sizes, etc.
+
+
# Remove the Greater London Authority duplicate
+csp_long2 %>%
+filter(authority !="Greater London Authority") %>%
+ggplot() +
+# Need to add a group to know what each line represents
+geom_line(aes(x = year, y = total_spend, group = ons_code,
+# OPTIONAL: colour by region to make it prettier!
+colour = region)) +
+facet_wrap( ~ region) +
+labs(x ="Year", y ="Total core spending power (millions)") +
+# Add theme_light to make the background a nicer colour
+theme_light() +
+# Rotate the x-axis labels to avoid overlap
+theme(axis.text.x.bottom =element_text(angle =45),
+# Remove the legend (not needed, we have labels on the facets)
+legend.position ="none")
+
+
+
+
+
+
+
+
+
+
+
Exercise 8
+
Create an RMarkdown file that creates a html report describing the trends in core spending power in English local authorities between 2015 and 2020. Your report should include:
+
+
A summary table of the total spending per year per region
+
A suitable visualisation showing how the total annual spending has changed over this period, compared between regions
+
A short interpretation of the table and visualisation
+
+
Note: You are not expected to be an expert in this data! Interpret these outputs as you would any other numeric variable measured over time.
+
+
+
Solutions
+
There are many different correct solutions to this exercise. All RMarkdown files should begin with a YAML header similar to the one below:
+
---
+title: "Core spending power in English local authorities, 2015 - 2020"
+author: Sophie Lee
+output: html_document
+---
+
Next, you may have a code chunk that sets up the global chunk options, loads any packages you needed, and loads the data that we will be using for the report. For example:
+
```{r setup, include = FALSE}
+# Set global chunk options to not show R code or messages
+knitr::opts_chunk$set(echo = FALSE, message = FALSE)
+
+# Load the tidyverse package
+library(tidyverse)
+
+# Load the long dataset
+csp_long2 <- read_csv("data/CSP_long_201520.csv")
+```
+
You may have began with an introduction using RMarkdown syntax:
+
# Introduction
+The following report will investigate the trends in core spending power
+across England between 2015 and 2020. All values are give in millions
+of pounds.
+
+The core spending power was made up of the following provisions:
+
+- Settlement funding assessment (SFA)
+- Compensation for under-indexing the business rates multipliers
+- council tax
+- New homes bonus
+- New homes bonus returned funding
+- Rural Services Delivery Grant (RSDG)
+
Followed by a summary table, created using summarise and displayed using kable:
+
# Total core spending power by region
+Below is a summary table containing the total core spending power per year
+per region, given in millions of £:
+
+```{r csp total summary table}
+csp_long2 %>%
+ group_by(region, year) %>%
+ summarise(min_spend = min(total_spend),
+ max_spend = max(total_spend),
+ median_spend = median(total_spend),
+ iqr_spend = IQR(total_spend)) %>%
+ ungroup() %>%
+ knitr::kable(.,
+ col.names = c("Region", "Year", "Minimum",
+ "Maximum", "Median", "IQR"))
+```
+
Then an additional code chunk producing a faceted line chart, similar to the one in Exercise 7:
+
+
+
+
+
\ No newline at end of file
diff --git a/_book/exercise_solutions_files/figure-html/aesthetic solution-1.png b/_book/exercise_solutions_files/figure-html/aesthetic solution-1.png
new file mode 100644
index 0000000..53389d2
Binary files /dev/null and b/_book/exercise_solutions_files/figure-html/aesthetic solution-1.png differ
diff --git a/_book/exercise_solutions_files/figure-html/region total spend line graph-1.png b/_book/exercise_solutions_files/figure-html/region total spend line graph-1.png
new file mode 100644
index 0000000..25d7885
Binary files /dev/null and b/_book/exercise_solutions_files/figure-html/region total spend line graph-1.png differ
diff --git a/_book/exercise_solutions_files/figure-html/scatter with line of best fit by region-1.png b/_book/exercise_solutions_files/figure-html/scatter with line of best fit by region-1.png
new file mode 100644
index 0000000..6ef173c
Binary files /dev/null and b/_book/exercise_solutions_files/figure-html/scatter with line of best fit by region-1.png differ
diff --git a/_book/exercise_solutions_files/figure-html/scatter with line of best fit-1.png b/_book/exercise_solutions_files/figure-html/scatter with line of best fit-1.png
new file mode 100644
index 0000000..27e9b56
Binary files /dev/null and b/_book/exercise_solutions_files/figure-html/scatter with line of best fit-1.png differ
diff --git a/_book/exercise_solutions_files/figure-html/sfa histogram-1.png b/_book/exercise_solutions_files/figure-html/sfa histogram-1.png
new file mode 100644
index 0000000..e3a4902
Binary files /dev/null and b/_book/exercise_solutions_files/figure-html/sfa histogram-1.png differ
diff --git a/_book/index.html b/_book/index.html
index 8cf954e..aae1bb3 100644
--- a/_book/index.html
+++ b/_book/index.html
@@ -7,7 +7,7 @@
-
+
Introduction to R with Tidyverse
icon and select ‘R script’.
