• Warm-up: data visualization
• Data Analysis by Hand
• Elements of Visualization
• Group Activity
• Grammar of Graphics

• Grammar of Graphics
• Intro to visualization in R
• Re-creating Visualizations using R

• Office Hours: Fridays, 1:30pm-3:00pm (Tyler)
• Tuesdays, 10:30am-12:00pm (Yao)

• Problem set and file naming
• GitHub/PSet Submissions
• PSet 3 and week 3 course site materials released shortly

##             group 1           group 2         group 3         group 4
## 1       Qin, Celine Mahoney, Brigette     Pham, Canon     Ong, Alyssa
## 2    Knowles, Genny      Barga, Jolie    Mendoza, Ava Bell, Mary Rose
## 3 Wolfenstein, Luci     Pacheco, Alex Batson, Anthony     Smith, Reid
##           group 5        group 6
## 1   Moore, Allana Devir, Lindsey
## 2                  Leahy, Olivia
## 3 Kang, Christine   Myoung, Sein

• Produce a visualization with the information on the handouts!
• You don’t need to use all the information or produce a finished product

##             group 1           group 2         group 3         group 4
## 1       Qin, Celine Mahoney, Brigette     Pham, Canon     Ong, Alyssa
## 2    Knowles, Genny      Barga, Jolie    Mendoza, Ava Bell, Mary Rose
## 3 Wolfenstein, Luci     Pacheco, Alex Batson, Anthony     Smith, Reid
##           group 5        group 6
## 1   Moore, Allana Devir, Lindsey
## 2                  Leahy, Olivia
## 3 Kang, Christine   Myoung, Sein

• In 1854, London suffered a cholera outbreak
• Unknown origin/cause
• Snow’s data analysis (and visualization) pointed to a water pump (Broad St Pump)
• Further analysis showed different rates of infection for those living in areas serviced by different water providers
• Interactive version here

• In 1869, Minard illustrated Napoleon’s March on Moscow (1812-1813)
• Combines geographical information, time, and population of the army

• Goal: illustrate the US color line at the Paris Exposition Universelle (1900)
• We might call this public sociology

• Highlighted both progress and inequality for Black Southern US Citizens
• Displayed Black Southerners as a group with commonalities to the future-oriented “thinking world,” in spite of position within segregated society

• How can color be used well (or poorly) to communicate social science?

• How can color be used well (or poorly) to communicate social science?
• Contrasts can be more (or less) visually appealing
• Color can points readers’ attention to or from certain areas
• Consider: existing associations with color

• How can scale be used well (or poorly) to communicate social science?

• How can scale be used well (or poorly) to communicate social science?
• Accurately reflect data
• Make comparisons easy, while holding viewer’s attention

• How can labels be used well (or poorly) to communicate social science?

• How can labels be used well (or poorly) to communicate social science?
• Readability
• Conciseness
• Placement

• How can comparisons be used well (or poorly) to communicate social science?

• How can comparisons be used well (or poorly) to communicate social science?
• Key comparisons should be easy for the viewer
• Details of comparison

• Re-visit your visualization with the information on the handouts!
• What do you want to communicate?
• How will you use color, scale, labels, and comparisons to make your point?

##             group 1           group 2         group 3         group 4
## 1       Qin, Celine Mahoney, Brigette     Pham, Canon     Ong, Alyssa
## 2    Knowles, Genny      Barga, Jolie    Mendoza, Ava Bell, Mary Rose
## 3 Wolfenstein, Luci     Pacheco, Alex Batson, Anthony     Smith, Reid
##           group 5        group 6
## 1   Moore, Allana Devir, Lindsey
## 2                  Leahy, Olivia
## 3 Kang, Christine   Myoung, Sein

• Finish your plots from Monday! And think about:
• What data would someone need to completely re-create your plot?
• What aesthetic instructions would someone need to completely re-create your plot?

• Grammar of Graphics
• Intro to visualization in R
• Re-creating Visualizations using R

• Office Hours: Fridays, 1:30pm-3:00pm (Tyler)
• Tuesdays, 10:30am-12:00pm (Yao)

• Project proposals are approaching (next week!)
• Campus events of interest:
• Sustainability and Data Science Conference (April 17, 9am-2:30pm)
• Empirical Work in the Age of AI (April 17, Oberdorff 1:30pm-5pm)

• The “grammar of graphics” is the gg in ggplot2
• It uses Object Oriented Design which is a system for how objects relate to one another
• This means its a little different than code we have written so far!
• Think: Data and Aesthetics

• Data is just our data (but we may need to manipulate it to be legible to ggplot2!)
• Aesthetics include colors, scales, labels, etc.

• We can try ggplot2 with the gapminder data
• What if we wanted more info on the gapminder data?

library(gapminder)

View(gapminder)

• Let’s try it out
• We can start with a blank canvas using ggplot

library(ggplot2)

# plot blank canvas
ggplot()

• Now let’s add to it
• We can input data and aesthetics
• Why don’t we see any data?

library(gapminder)

# add data and aesthetics
ggplot(data = gapminder, mapping = aes(x = gdpPercap, y = lifeExp))

• ggplot2 works in layers!
• To see our data, we need to add a geom_ layer

• The + sign adds layers to our ggplot
• We could try geom_point
• What about another geom?

# create a plot
ggplot(data = gapminder,
            mapping = aes(x = gdpPercap, y = lifeExp))+
  geom_point()

• There are multiple ways to do this
• When/why would we want to assign an object (p) to our plot?

# create the plot
p <- ggplot(data = gapminder,
            mapping = aes(x = gdpPercap, y = lifeExp))

# plot it
p + geom_point()

• We can add color within the geom_point function
• First: manually set

# create the plot
ggplot(data = gapminder,
        mapping = aes(x = gdpPercap, y = lifeExp))+
  geom_point(col = "purple")

• We can add color within the geom_point function
• First: manually set
• Second: within aesthetics

# create the plot
ggplot(data = gapminder,
        mapping = aes(x = gdpPercap, y = lifeExp,
        col = continent))+
  geom_point()

• Let’s try it!
• Set the color to vary by year
• Manually set alpha to 0.5
• Set the size to vary by pop
• Try changing other aesthetics! Can you make a different type of plot?

# create the plot
ggplot(data = gapminder,
        mapping = aes(x = gdpPercap, y = lifeExp,
        col = continent))+
  geom_point()

# create a dataframe (note that : returns a sequence)
df <- data.frame(year = 2000:2020, pop = 40:60)

# look at the first 6 rows
head(df)

##   year pop
## 1 2000  40
## 2 2001  41
## 3 2002  42
## 4 2003  43
## 5 2004  44
## 6 2005  45

library(dplyr)

# function to get from millions to people
millions_to_people <- function(m){
  p <- m/1000000
  return(p)
}

# run function to create new variable
df %>%
  mutate(pop_p = millions_to_people(pop)) %>%
  head() 

##   year pop   pop_p
## 1 2000  40 4.0e-05
## 2 2001  41 4.1e-05
## 3 2002  42 4.2e-05
## 4 2003  43 4.3e-05
## 5 2004  44 4.4e-05
## 6 2005  45 4.5e-05

• On your own try creating a dataframe and dividing pop by 100

# create a dataframe
df <- data.frame(year = 2000:2020, pop = 40:60)


# create your function here
 <- function(x){
   
 }

# modify your dataframe here
df

Instead of this:

fourth_function(third_function(second_function(first_function(x))))

We do this:

x %>%
  first_function() %>% 
  second_function() %>% 
  third_function() %>%
  fourth_function()

• How would we permanently change df?

# run function on to create new variable
df %>%
  mutate(pop_p = millions_to_people(pop)) %>%
  head() 

##   year pop   pop_p
## 1 2000  40 4.0e-05
## 2 2001  41 4.1e-05
## 3 2002  42 4.2e-05
## 4 2003  43 4.3e-05
## 5 2004  44 4.4e-05
## 6 2005  45 4.5e-05

• We can permanently change an object with the magrittr pipe (%<>%)

library(magrittr)

# we use the magrittr pipe
df %<>%
  mutate(pop_p = millions_to_people(pop))

# check that it worked
head(df)

• Try using the magrittr pipe!
• Is there a way to do the same thing with the <- and %>% operators?

library(magrittr)

# we use the magrittr pipe
df %<>%
  mutate(pop_p = millions_to_people(pop))

• We can filter to observations of interest

library(readr)
library(RCurl)

# input url from github
url <- getURL("https://raw.githubusercontent.com/tylermcdaniel/soc10/refs/heads/deployment/Data/ca_dem_table.csv")

# download wikipedia data (from github)
ca_dem_table <- read_csv(url)

• What if we want to divide 2000_num by the total for that year (33871648) and then remove the “total” row?

ca_dem_table %<>%

• In Problem Set 3, you can create a ggplot with the California demographic data
• Try reshaping the data with magrittr

• Start with data and aesthetics
• Add layers to your plot
• Consider which elements you want to manually set vs. vary with aesthetics

• Project proposals are approaching (next week!)
• Campus events of interest:
• Sustainability and Data Science Conference (April 17, 9am-2:30pm)
• Empirical Work in the Age of AI (April 17, Oberdorff 1:30pm-5pm)