Lecture 7
Adam Lucas
Source file ⇒ 2017-lec5.Rmd
Announcements
- thank you for you feedback!
- HW 2 is out!
Today:
- Review ideas of from DC Chapter 6 (Frames, Glyphs, and other components of graphs)
- DC Chapter 8 ggplot2
- DC Chapter 7 and 9 (Wrangling and Data Verbs)
Chapter 6
Volcabulary:
roads <- read.csv(file="http://tiny.cc/dcf/table-6-2.csv")
head(roads)## country gdp educ roadways net_users
## 1 Albania 9383.46 3.3 0.63 >35%
## 2 Algeria 7335.03 4.3 0.05 >5%
## 3 Angola 6904.82 3.5 0.04 >0%
## 4 Anguilla 10903.89 2.8 1.92 >15%
## 5 Antigua and Barbuda 17635.14 2.4 2.64 >60%
## 6 Argentina 17920.07 6.3 0.08 >15%
In class exercise: What are these ggplot2 components for the data table and plot above?
Frame= The space for drawing gyphs.
Glyph= A symbol or a marking in a frame. —rectangular region Aesthetics= The properties of a glyph related to a variable in the dataset. —x,y,size
Scales= the mapping between property of glyph and variable in dataset —-x=roadways, y=gdp, size=educ
Guide= a display of scale. —tick marks x and y axis, facet guide, legend
Facets= Multiple side by side graphs used to display levels of a categorical variable. —-internet usage is a facet Layers= multiple glyphs —only one
Here is were to find out the details of the ggplot2 components:
http://docs.ggplot2.org/current/
Chapter 8: Graphics and their Grammar
ggplot2 is a grpahics package that uses the components of graphs (i.e. glyphs, aestetics, frames, scales, layers) –called the grammar of graphics.
We call glyphs geoms now.
examples
Here is the data table mosaicData::CPS85:
data(CPS85,package="mosaicData")
head(CPS85)## wage educ race sex hispanic south married exper union age sector
## 1 9.0 10 W M NH NS Married 27 Not 43 const
## 2 5.5 12 W M NH NS Married 20 Not 38 sales
## 3 3.8 12 W F NH NS Single 4 Not 22 sales
## 4 10.5 12 W F NH NS Married 29 Not 47 clerical
## 5 15.0 12 W M NH NS Married 40 Union 58 const
## 6 9.0 16 W F NH NS Married 27 Not 49 clericalframe <- CPS85 %>% ggplot(aes(x=age,y=wage))
frame + geom_point()
frame <- CPS85 %>% ggplot(aes(x=age,y=wage))
frame + geom_point(aes(shape=sex))
frame <- CPS85 %>% ggplot(aes(x=age,y=wage))
frame + geom_point(aes(shape=sex)) + facet_grid(married ~ .)
frame <- CPS85 %>% ggplot(aes(x=age,y=wage))
frame + geom_point(aes(shape=married)) + ylim(0,30)
Here are my two favorite resources for ggplot2:
In class exercise
Please make from data table CPS85.
Chapter 7 Wrangling and Data Verbs
Rarely is a data table glyph ready. We usually need to wrangle the data. A data verb is a function that takes a data table as an input and produces a data table as an output.
summarise() and group_by()
Two of the most useful Data Verbs often used together are summarise() and group_by():
summarise() turns multiple cases into a single case using reduction formulas such as n() or sum() or mean().
For example:
head(DataComputing::BabyNames)## name sex count year
## 1 Mary F 7065 1880
## 2 Anna F 2604 1880
## 3 Emma F 2003 1880
## 4 Elizabeth F 1939 1880
## 5 Minnie F 1746 1880
## 6 Margaret F 1578 1880BabyNames %>% summarise(num_cases=n()) #gives the number of rows.## num_cases
## 1 1792091here n() is a reduction function that only works inside of summarise()
Instead of using summarise(num_cases= n()) you could use nrow() or tally().
BabyNames %>% nrow()## [1] 1792091BabyNames %>% tally()## n
## 1 1792091tally() is a data verb (since it outputs a data table) but nrow() isn’t since it outputs a number.
BabyNames %>% summarise(average=mean(count)) #gives the average of counts## average
## 1 186.0496mean() or sum() are reduction functions that takes a variable (for example count) as an argument.
If we want to summarise by name we would use group_by() and summarise() together.
BabyNames %>%
group_by(name) %>%
summarise(num_cases=n()) ## # A tibble: 92,600 × 2
## name num_cases
## <chr> <int>
## 1 Aaban 6
## 2 Aabha 2
## 3 Aabid 1
## 4 Aabriella 1
## 5 Aadam 22
## 6 Aadan 8
## 7 Aadarsh 13
## 8 Aaden 14
## 9 Aadesh 3
## 10 Aadhav 7
## # ... with 92,590 more rowsIf we want to divide the name into subgroups according to the sex of the baby, let group_by() have two arguements name and sex (order matters).
BabyNames %>%
group_by(name, sex) %>%
summarise(num_cases=n(), sum_cases=sum(count)) ## Source: local data frame [102,690 x 4]
## Groups: name [?]
##
## name sex num_cases sum_cases
## <chr> <chr> <int> <int>
## 1 Aaban M 6 56
## 2 Aabha F 2 12
## 3 Aabid M 1 5
## 4 Aabriella F 1 5
## 5 Aadam M 22 177
## 6 Aadan M 8 104
## 7 Aadarsh M 13 140
## 8 Aaden F 1 5
## 9 Aaden M 13 3677
## 10 Aadesh M 3 15
## # ... with 102,680 more rowsNotice that the table now has 4 colunmns.
What is the 3 most popular names in BabyNames?
BabyNames %>%
group_by(name) %>%
summarise(tot=sum(count)) %>%
arrange(desc(tot)) %>%
head(3)## # A tibble: 3 × 2
## name tot
## <chr> <int>
## 1 James 5114325
## 2 John 5095590
## 3 Robert 4809858In class exercises:
I suggest you open this in another browser (not in rpubs)
http://gandalf.berkeley.edu:3838/alucas/Chapter-07-collection/
Suppose we want to make a bar chart ranking the candidates by how many precincts they are in first place.
#not glyph ready
head(DataComputing::Minneapolis2013)## Precinct First Second Third Ward
## 1 P-10 BETSY HODGES undervote undervote W-7
## 2 P-06 BOB FINE MARK ANDREW undervote W-10
## 3 P-09 KURTIS W. HANNA BOB FINE MIKE GOULD W-10
## 4 P-05 BETSY HODGES DON SAMUELS undervote W-13
## 5 P-01 DON SAMUELS undervote undervote W-5
## 6 P-04 undervote undervote undervote W-6FirstPlaceTally <- Minneapolis2013 %>%
rename(candidate=First) %>%
group_by(candidate) %>%
summarise(total=n()) %>%
arrange( desc(total))
#glyph ready
FirstPlaceTally## # A tibble: 38 × 2
## candidate total
## <chr> <int>
## 1 BETSY HODGES 28935
## 2 MARK ANDREW 19584
## 3 DON SAMUELS 8335
## 4 CAM WINTON 7511
## 5 JACKIE CHERRYHOMES 3524
## 6 BOB FINE 2094
## 7 DAN COHEN 1798
## 8 STEPHANIE WOODRUFF 1010
## 9 MARK V ANDERSON 975
## 10 undervote 834
## # ... with 28 more rowsIn Class exercises
- Suppose want to order the Precincts according to which has the most number of ballots cast
Minneapolis2013 %>%
group_by(Precinct) %>%
summarise(count=n()) %>% # n() finds how many cases there are
arrange(desc(count))## # A tibble: 16 × 2
## Precinct count
## <chr> <int>
## 1 P-06 9711
## 2 P-02 9551
## 3 P-08 9430
## 4 P-03 8703
## 5 P-05 8490
## 6 P-07 8104
## 7 P-04 7753
## 8 P-01 7301
## 9 P-09 5342
## 10 P-10 1561
## 11 P-04D 852
## 12 P-02D 822
## 13 P-05A 742
## 14 P-03A 730
## 15 P-01C 505
## 16 P-6C 504#or
Minneapolis2013 %>%
group_by(Precinct) %>%
tally(sort=TRUE) ## # A tibble: 16 × 2
## Precinct n
## <chr> <int>
## 1 P-06 9711
## 2 P-02 9551
## 3 P-08 9430
## 4 P-03 8703
## 5 P-05 8490
## 6 P-07 8104
## 7 P-04 7753
## 8 P-01 7301
## 9 P-09 5342
## 10 P-10 1561
## 11 P-04D 852
## 12 P-02D 822
## 13 P-05A 742
## 14 P-03A 730
## 15 P-01C 505
## 16 P-6C 504- What if want to order Precincts with the most number of ballots cast having “BETSY HODGES” as First?
Minneapolis2013 %>%
filter(First =="BETSY HODGES") %>%
group_by(Precinct) %>%
tally(sort=TRUE)## # A tibble: 16 × 2
## Precinct n
## <chr> <int>
## 1 P-06 3762
## 2 P-02 3739
## 3 P-08 3480
## 4 P-07 3073
## 5 P-05 2895
## 6 P-01 2793
## 7 P-03 2663
## 8 P-04 2571
## 9 P-09 1943
## 10 P-10 486
## 11 P-02D 326
## 12 P-04D 319
## 13 P-05A 303
## 14 P-03A 295
## 15 P-01C 161
## 16 P-6C 126- How many votes are marked “BETSY HODGES” in the First and Second choice selection?
Minneapolis2013 %>%
filter(Second == "BETSY HODGES", First == "BETSY HODGES") %>%
tally() #could also use nrow here## n
## 1 222Minneapolis2013 %>%
group_by(First,Second) %>%
tally() %>%
filter(First=="BETSY HODGES", Second=="BETSY HODGES")## Source: local data frame [1 x 3]
## Groups: First [1]
##
## First Second n
## <chr> <chr> <int>
## 1 BETSY HODGES BETSY HODGES 222