Data Visualization¶

Getting Started¶

Please follow these steps to install necessary software before the session:

• Download and install R
• Download and install RStudio
• Download and install Tableau Public. You may be asked to sign up with your email address.

Exercises¶

In teams of 2, select one of the following 10 datasets to use. Your task is to:

1. Get a basic understanding of the data set: use the internet, make summary statistics, etc.
2. Explore the data and think about some questions you might want to answer through visual representation of the data
3. Choose an audience that you will create a visualization and a report for: government officials, a private sector partner, your PIs, a classroom of college students, etc.
4. Choose what tools you will choose to produce the visualization (Raw, Tableau, R, STATA, etc.)
5. Get to work and produce an interesting graphic that (i) clearly highlights what you are investigating (in terms of variables, measures, dimensions, etc.) (ii) is easily digestible and (iii) is supported by a description substantiating your findings (even if there are none)
6. Email us your final visualization (PNG / URL). We will put each of these visualizations up, and you will get to quickly explain your intent and findings (if any).

ggplot2¶

We will use the mtcars dataset, which comes pre-loaded in the R session. Explore the dataset by executing str(mtcars) and summary(mtcars) in the R console.

Load the ggplot2 package using:

library(ggplot2)

plot(mtcars$hp,mtcars$mpg)

scatterplot <- ggplot(mtcars, aes(x = hp,y = mpg)) +
        geom_point()

scatterplot <- scatterplot +
  labs(title = "MPG vs HP",
       x = "HP",
       y = "Miles Per Gallon")

mtcars$gear <- as.factor(mtcars$gear)

scatterplot <- ggplot(mtcars, aes(x = mpg,y = hp)) +
  geom_point(aes(colour = mtcars$gear)) +
  geom_smooth(method = "lm",se = FALSE) +
  scale_colour_discrete(name = 'Gear',
                        breaks = c('3','4','5'),
                        labels = c('Low','Medium','High')) +
  labs(title = "MPG vs HP",
       x = "HP",
       y = "Miles Per Gallon")

ggplot(mtcars, aes(x = mpg)) +
  geom_histogram(binwidth = 5)

ggplot(mtcars, aes(x = mpg)) +
  geom_histogram(binwidth = 5,
                 colour = "black",
                 fill = "white") +
  labs(title = "Distribution - Miles Per Gallon",
       x = "Miles Per Gallon",
       y = "Count")

Saving your graphs¶

?ggsave gives you a description of the parameters that the function takes, and their defaults:

ggsave(filename, plot = last_plot(), device = NULL, path = NULL,
        scale = 1, width = NA, height = NA, units = c("in", "cm", "mm"),
        dpi = 300, limitsize = TRUE, ...)

To save our scatterplot, we might want to use the following parameters:

ggsave(filename = 'FULL_FILE_PATH.FILE_TYPE',
        plot = scatterplot,
        width = 6,height = 5.4)

FILE_TYPE can be tex, pdf, jpeg, png, svg and others (refer documentation by using ?ggsave).

Output

ggvis¶

• Data: Infant Mortality Rates (India and major states) by Sex - 2001 to 2012
• Data Source: NRHM
• Download Data

• Install / load required packages using:

  # Check to see if packages are installed. Install them if they are not, then load them into the R session.
  
  ipak <- function(pkg){
    new.pkg <- pkg[!(pkg %in% installed.packages()[, "Package"])]
    if (length(new.pkg))
      install.packages(new.pkg, dependencies = TRUE)
    sapply(pkg, require, character.only = TRUE)
  }
  
  packages <- c("dplyr", "xts", "lubridate", "tidyr", "ggvis")
  ipak(packages)
  

• Set file paths:

  #Set file paths
  
  ROOT <- "FULL_FILE_PATH"   #Note that you need to replace \ with /
  DATA <- paste0(ROOT,"data/")
  OUTPUT <- paste0(ROOT,"output/")
  #---------------------------------------------------------------------
  

• Read CSV file:

  imrData <- read.csv(paste0(DATA,'IMRsex2001-2012.csv'))
  

• Prepare data for analysis:

  imrData <- imrData %>%
    select(state = India.States.Uts,
           X2001.Total:X2012.Female) %>%  #Keep selected variables
    gather(yearSex,imr,X2001.Total:X2012.Female) %>%    #Reshape data
    separate(col = yearSex,
             into = c('year','sex'),
             sep = "\\.",
             remove = TRUE) %>%         #Split yearSex column into year and sex columns
    filter(sex != 'Total') %>%          #Remove observations of category 'Total'
    mutate(year = as.Date(paste0(substr(year,start = 2,stop = 5),'-12-31')))   #Remove 'X' at the beginning of year, convert variable into date.
  

• Plot bar graph:

  imrData %>%
    filter(state != 'INDIA',year == max(year)) %>%
    mutate(state = as.character(state)) %>%
    ggvis(x = ~state,y = ~imr) %>%
    layer_bars() %>%
    add_axis("x",title = "",
             properties = axis_props(labels = list(angle = 270,
                                                   dy = -5,
                                                   align = "right")))
  

Output

htmlwidgets¶

htmlwidgets allows us to use the power of Javascript-based visualization libraries from within R, to produce some fantastic visualizations.

Tips for working with R¶

Dropbox Paths

If you’re on a project that uses Dropbox for file sharing, you can set Dropbox paths (assuming default directory) using the following snippet of code.

Tableau Public¶

• Data: Total Sex Ratio (males / hundred females)
• Data Source: UN Population Division (posted on Gapminder)
• Download Data

Tableau Public is quite intuitive and easy to use. We will walk through each step to reach this visualization: