Why learn R?

Environments and systems

R is:

  • A statistical progamming language
  • Open source
  • An essential toolkit for data science and modern analysis
    • A GIS
    • A modelling toolkit
    • A machine learning toolkit
    • A time series analysis toolkit
    • A suite of data wrangling tools
    • A suite of high quality visualisation tools
    • A set of tools for creating interactive vizzes and dashboards
    • A powerful text mining toolkit
    • A set of tools for undertaking reproducible open science
    • A tool for creating and building analytical and data services

Getting up and running (quickly)

Getting data into R

  • Can import data is any format e.g.
    • csv
      • readr::read_csv()
      • data.table::fread() - fast read for large files
    • xlsx
      • readxl::read_excel()
    • ods
      • ODS::read_ods()
    • .dta / .sav
      • haven::read_sav() / haven::read_spss()
      • haven::read_dta()
    • text files
      • readr::read_delim()
  • R also has its own data format .rds which is fast and highly compressed
library(pacman)
p_load(tidyverse)

data <- read_csv("https://coronavirus.data.gov.uk/api/v1/data?filters=areaType=nation;areaName=England&structure=%7B%22areaType%22:%22areaType%22,%22areaName%22:%22areaName%22,%22areaCode%22:%22areaCode%22,%22date%22:%22date%22,%22uniqueCasePositivityBySpecimenDateRollingSum%22:%22uniqueCasePositivityBySpecimenDateRollingSum%22,%22uniquePeopleTestedBySpecimenDateRollingSum%22:%22uniquePeopleTestedBySpecimenDateRollingSum%22%7D&format=csv", show_col_types = F)
head(data)
## # A tibble: 6 × 6
##   areaType areaName areaCode  date       uniqueCasePositivityB… uniquePeopleTes…
##   <chr>    <chr>    <chr>     <date>                      <dbl>            <dbl>
## 1 nation   England  E92000001 2022-01-21                   16.9          2122387
## 2 nation   England  E92000001 2022-01-20                   16.9          2148550
## 3 nation   England  E92000001 2022-01-19                   17.2          2171736
## 4 nation   England  E92000001 2022-01-18                   17.5          2192894
## 5 nation   England  E92000001 2022-01-17                   18.1          2257634
## 6 nation   England  E92000001 2022-01-16                   19            2375188

Core tidy functions

dplyr

Data wrangling package with a set of key functions (verbs)

filter - operations to select rows on the basis of specified criteria

data %>%
  filter(date == "2022-01-01")
## # A tibble: 1 × 6
##   areaType areaName areaCode  date       uniqueCasePositivityB… uniquePeopleTes…
##   <chr>    <chr>    <chr>     <date>                      <dbl>            <dbl>
## 1 nation   England  E92000001 2022-01-01                   31.2          3306581

select - select columns

data %>%
  select(date, uniqueCasePositivityBySpecimenDateRollingSum) %>%
  filter(date >= "2022-01-01")
## # A tibble: 21 × 2
##    date       uniqueCasePositivityBySpecimenDateRollingSum
##    <date>                                            <dbl>
##  1 2022-01-21                                         16.9
##  2 2022-01-20                                         16.9
##  3 2022-01-19                                         17.2
##  4 2022-01-18                                         17.5
##  5 2022-01-17                                         18.1
##  6 2022-01-16                                         19  
##  7 2022-01-15                                         19.8
##  8 2022-01-14                                         20.7
##  9 2022-01-13                                         22  
## 10 2022-01-12                                         23.8
## # … with 11 more rows

mutate - add a new column

data %>%
  mutate(cases = uniqueCasePositivityBySpecimenDateRollingSum / uniqueCasePositivityBySpecimenDateRollingSum[1]) %>%
  select(-areaType) %>%
  head()
## # A tibble: 6 × 6
##   areaName areaCode  date       uniqueCasePositivityBySp… uniquePeopleTes… cases
##   <chr>    <chr>     <date>                         <dbl>            <dbl> <dbl>
## 1 England  E92000001 2022-01-21                      16.9          2122387  1   
## 2 England  E92000001 2022-01-20                      16.9          2148550  1   
## 3 England  E92000001 2022-01-19                      17.2          2171736  1.02
## 4 England  E92000001 2022-01-18                      17.5          2192894  1.04
## 5 England  E92000001 2022-01-17                      18.1          2257634  1.07
## 6 England  E92000001 2022-01-16                      19            2375188  1.12

Plotting with ggplot2

data %>%
  mutate(cases = uniqueCasePositivityBySpecimenDateRollingSum / uniqueCasePositivityBySpecimenDateRollingSum[1]) %>%
  ggplot(aes(date, cases)) +
  geom_line(lty = "dotted") +
  geom_smooth(method = "gam", se = FALSE) +
  theme_bw()
## `geom_smooth()` using formula 'y ~ s(x, bs = "cs")'
## Warning: Removed 1 rows containing non-finite values (stat_smooth).

NE Data

devtools::install_github("ropensci/fingertipsR")
## Skipping install of 'fingertipsR' from a github remote, the SHA1 (3d4ea2a7) has not changed since last install.
##   Use `force = TRUE` to force installation
library(fingertipsR); library(tidyverse)

#fingertipsR::select_indicators()

smoking <- fingertips_data(IndicatorID = c(92443, 93085, 93553, 93579, 
                                           93573, 93798, 91547, 92304), 
                           AreaTypeID = 401)


smoking_ne <- smoking %>%
  filter(ParentCode == "E12000001")

Smoking prevalence trend in NE districts

Code

smoking_ne %>%
  filter(str_detect(IndicatorName, "[Pp]revalence"), Age == "18+ yrs", 
  Sex == "Persons") %>%
  ggplot(aes(round(TimeperiodSortable/10^4, 0), Value, 
  colour = IndicatorName)) +
  geom_line(aes(group = IndicatorName)) +
  geom_point() +
  geom_linerange(aes(ymin = LowerCI95.0limit, ymax = UpperCI95.0limit)) +
  facet_wrap(~AreaName, 
             nrow = 3) +
  theme(legend.position = "bottom", 
        axis.text.x = element_text(angle = 0, hjust =1)) +
  labs(x = "Year", 
       y = "Estimate (%)") +
  scale_x_continuous(n.breaks = 6)