Visualization

Workshop date: 2018-04-28
Author: Jon Pipitone, 2018

This workshop is an interactive exploration of ggplot, the plotting library that is part of the tidyverse.

Big Ideas

Here are the big takeaways I’d like you to leave with from this workshop:

  • Restructure your data to make plotting easier

    For certain kinds of visualizations, it helps to first reorganize your data. For instance, rather than columns for each kind of measurement, we use a new row.

    Time    Measurement   Value
11:11   HR            65 
11:11   SpO2          99
11:12   HR            110
11:12   SpO2          87
...

  • Pipe/filter your data to ggplot

    You can pipe your data directly to ggplot but massage it along the way using dplyr verbs to get just want you want to plot.

  • Create charts by adding layers

    With ggplot, you add layers or override default attributes of your plot to specify what should be present.

Learning objectives

  • Review RStudio, and parts of the IDE: console, editor, heap
  • Install and load libraries using install.packages(), library()
  • Review the “pipe and filter” model of dplyr
  • Review dplyr verbs: select, mutate, filter
  • Introduce piping tibbles to ggplot()
  • Use “aesthetics”, aes(), to specify which data attributes to use in a plot
  • Create different visualizations of the data with geom_point, geom_line, geom_bar, geom_histogram, geom_boxplot
  • Use dplyr’s gather function to “tidy” your dataset
  • Create arrays of plots using facet_wrap
  • Customizing labels and other accessories on plots
  • Exporting/saving plots using ggsave()

Tasks

Set up

  1. Install RStudio and R
  2. Install the tidyverse package using the install.packages function, and load it using the library() function: 
    install.packages('tidyverse')     # this may take some time
library(tidyverse)
  3. Download the dataset: 
     download.file(
     'https://pipitone.github.io/qmed-computes/assets/workshop-data/ggplot-timeseries.csv',
     destfile = 'data.csv')

Script

Since this workshop was interactive, there was no set of specific tasks to follow. Here’s a script from one of the sessions that covers some of what we did:


install.packages('tidyverse')
library(tidyverse)

download.file(
  url = 'https://pipitone.github.io/qmed-computes/assets/workshop-data/ggplot-timeseries.csv',
  destfile = 'data.csv')

# Review some data wrangling

data = read_csv('data.csv')

spo2_table = data %>% 
  select(Time, SpO2)

spo2_table = data %>% 
  select(-Time)

data2 = data %>% 
  select(Time, HR, SpO2) %>%
  filter(HR < 100) %>%
  mutate(tachy = HR > 100, 
         test = SpO2 / 100)

write_csv(data2, 'data2.csv')

# make a simple scatter/line plot
data %>%
  filter(HR < 120) %>%
  ggplot(aes(x = Time, y = HR)) +
    geom_line() + 
    geom_point() +
    labs(
      title = "Title", 
      subtitle = "subtitle", 
      x = "Time (s)",
      caption = "Source: David Wood"
    ) + 
    theme_minimal()

# use gather() to rejig our data to make it possible to plot by measurement
data %>%
  mutate(tachy = HR > 100) %>%
  gather(Measurement, Value, -Time, -tachy) %>%
  ggplot(aes(x = Time, y = Value, colour = Measurement)) + 
    geom_point() + 
    geom_line() + 
    facet_wrap(~ Measurement)

Resources