Data Summarization Lab Key
Data used
Bike Lanes Dataset: BikeBaltimore is the Department of Transportation’s bike program. The data is from http://data.baltimorecity.gov/Transportation/Bike-Lanes/xzfj-gyms
You can Download as a CSV in your current working directory. Note its also available at: http://jhudatascience.org/intro_to_R_class/data/Bike_Lanes.csv
library(readr)
library(dplyr)
library(tidyverse)
library(jhur)
bike = read_csv(
"http://jhudatascience.org/intro_to_R_class/data/Bike_Lanes.csv")or use
library(jhur)
bike <- read_bike()## Rows: 1631 Columns: 9## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (6): subType, name, block, type, project, route
## dbl (3): numLanes, length, dateInstalled##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.Part 1
- How many bike “lanes” are currently in Baltimore? You can assume each observation/row is a different bike “lane”. (hint: how do you get the number of rows of a data set? You can use
dim()ornrow()or another function).
nrow(bike)## [1] 1631dim(bike)## [1] 1631 9bike %>%
nrow()## [1] 1631- How many feet of bike “lanes” are currently in Baltimore, based on the
lengthcolumn? (usesum())
sum(bike$length)## [1] 439447.6sum(bike %>% pull(length))## [1] 439447.6- Summarize the data to get the
maxandminoflengthusing thesummarizefunction.
bike %>% summarize(max = max(length),
min = min(length))## # A tibble: 1 x 2
## max min
## <dbl> <dbl>
## 1 3749. 0Part 2
- How many types of bike lanes are there? (Hints:
unique,table, orbike %>% count()on the column namedtype).
# So many ways!!
bike %>% count(type)## # A tibble: 8 x 2
## type n
## <chr> <int>
## 1 BIKE BOULEVARD 49
## 2 BIKE LANE 621
## 3 CONTRAFLOW 13
## 4 SHARED BUS BIKE 39
## 5 SHARROW 589
## 6 SIDEPATH 7
## 7 SIGNED ROUTE 304
## 8 <NA> 9bike %>% pull(type) %>% table()## .
## BIKE BOULEVARD BIKE LANE CONTRAFLOW SHARED BUS BIKE SHARROW
## 49 621 13 39 589
## SIDEPATH SIGNED ROUTE
## 7 304unique(bike %>% pull(type))## [1] "BIKE BOULEVARD" "SIDEPATH" "SIGNED ROUTE" "BIKE LANE"
## [5] "SHARROW" NA "CONTRAFLOW" "SHARED BUS BIKE"table(bike$type, useNA = "ifany")##
## BIKE BOULEVARD BIKE LANE CONTRAFLOW SHARED BUS BIKE SHARROW
## 49 621 13 39 589
## SIDEPATH SIGNED ROUTE <NA>
## 7 304 9unique(bike$type)## [1] "BIKE BOULEVARD" "SIDEPATH" "SIGNED ROUTE" "BIKE LANE"
## [5] "SHARROW" NA "CONTRAFLOW" "SHARED BUS BIKE"length(table(bike$type))## [1] 7length(unique(bike$type))## [1] 8is.na(unique(bike$type))## [1] FALSE FALSE FALSE FALSE FALSE TRUE FALSE FALSEbike %>%
group_by(type) %>%
summarize(count = n())## # A tibble: 8 x 2
## type count
## <chr> <int>
## 1 BIKE BOULEVARD 49
## 2 BIKE LANE 621
## 3 CONTRAFLOW 13
## 4 SHARED BUS BIKE 39
## 5 SHARROW 589
## 6 SIDEPATH 7
## 7 SIGNED ROUTE 304
## 8 <NA> 9- Which bike lane
typehas (a) the most number of and (b) longest average bike lane length? (Hint:group_byandsummarize). In your summarized output, make sure you call the new summarized average bike lane length variable (column name) “mean”. In other words, the head of your output should look like:
# A tibble:
type number_of_rows mean
<chr> <int> <dbl>
1 BIKE BOULEVARD 49 197.
...bike %>%
group_by(type) %>%
summarize(number_of_rows = n(),
mean = mean(length))## # A tibble: 8 x 3
## type number_of_rows mean
## <chr> <int> <dbl>
## 1 BIKE BOULEVARD 49 197.
## 2 BIKE LANE 621 300.
## 3 CONTRAFLOW 13 136.
## 4 SHARED BUS BIKE 39 277.
## 5 SHARROW 589 244.
## 6 SIDEPATH 7 666.
## 7 SIGNED ROUTE 304 264.
## 8 <NA> 9 260.- Take your code from the above question and do the following:
- Add another pipe (
%>%) - Add the
arrange()to sort the output by the summarized column “mean”.
bike %>%
group_by(type) %>%
summarize(number_of_rows = n(),
mean = mean(length)) %>%
arrange(mean)## # A tibble: 8 x 3
## type number_of_rows mean
## <chr> <int> <dbl>
## 1 CONTRAFLOW 13 136.
## 2 BIKE BOULEVARD 49 197.
## 3 SHARROW 589 244.
## 4 <NA> 9 260.
## 5 SIGNED ROUTE 304 264.
## 6 SHARED BUS BIKE 39 277.
## 7 BIKE LANE 621 300.
## 8 SIDEPATH 7 666.- Make a new column using
mutate. This new column should be different for each year (dateInstalled), and indicate the total sum number of lanes (numLanes) built in that year. Call this new columnyear_totaland make sure to reassign the dataset. (hint: usegroup_byfirst)
bike <- bike %>%
group_by(dateInstalled) %>%
mutate(year_total = sum(numLanes, na.rm = TRUE))- How does your data from above change if you also group by
type? Ungroup your data when you are done.
bike %>%
group_by(dateInstalled, type) %>%
mutate(year_total = sum(numLanes, na.rm = TRUE))## # A tibble: 1,631 x 10
## # Groups: dateInstalled, type [28]
## subType name block type numLanes project route length dateInstalled
## <chr> <chr> <chr> <chr> <dbl> <chr> <chr> <dbl> <dbl>
## 1 <NA> <NA> <NA> BIKE B… 1 GUILFOR… <NA> 436. 0
## 2 <NA> <NA> <NA> SIDEPA… 1 <NA> NORT… 1025. 2010
## 3 <NA> <NA> <NA> SIGNED… 1 SOUTHEA… <NA> 3749. 2010
## 4 <NA> HUNTIN… <NA> SIDEPA… 1 <NA> <NA> 0 0
## 5 STCLN EDMOND… 5300 BL… BIKE L… 1 OPERATI… <NA> 181. 2011
## 6 STRALY WINSTO… 1200 BL… SIGNED… 2 COLLEGE… COLL… 148. 2007
## 7 STRALY WINSTO… 1200 BL… SIGNED… 2 COLLEGE… COLL… 366. 2007
## 8 STRALY WINSTO… 1200 BL… SIGNED… 2 COLLEGE… COLL… 262. 2007
## 9 STRPRD <NA> <NA> BIKE L… 1 MAINTEN… <NA> 696. 2009
## 10 STRPRD <NA> <NA> SHARROW 1 COLLEGE… <NA> 43.1 2007
## # … with 1,621 more rows, and 1 more variable: year_total <dbl>bike <- ungroup(bike)Part 2
- Create a histogram for the
lengthvariable in thebikedataset. Try playing with thebreaks=argument.
hist( pull(bike, length), breaks = 100)
- Create a scatterplot with
dateInstalledon the x axis andlengthon the y axis.
plot( pull(bike, dateInstalled), pull(bike, length) )
Bonus
A. Summarize the bike data to get the mean of length and dateInstalled. Do this three ways: 1) with summarize, 2) with summarize and across, and 3) with colMeans().
bike %>% summarize( mean_length = mean(length, na.rm = TRUE),
mean_date = mean(dateInstalled, na.rm = TRUE))## # A tibble: 1 x 2
## mean_length mean_date
## <dbl> <dbl>
## 1 269. 1854.bike %>%
summarize(across( c(length, dateInstalled), ~ mean(.x, na.rm = TRUE)))## # A tibble: 1 x 2
## length dateInstalled
## <dbl> <dbl>
## 1 269. 1854.bike %>% select(length, dateInstalled) %>% colMeans()## length dateInstalled
## 269.4344 1853.9454You should have gotten a mean date sometime in the 1800s - that doesn’t make much sense! Hypothesize why the average date is a date from before bike lanes were being built in Baltimore.
There are probably some zeros or other incorrect low values in the data.
B. Change any zeros in bike$dateInstalled to NA using mutate. For the curious,ifelse() in R works takes the same arguments as the “IF” function in Excel!
bike <- bike %>%
mutate(dateInstalled = ifelse(dateInstalled == 0, NA, dateInstalled))
bike$dateInstalled[bike$dateInstalled == "0"] <- NA
# How to find NAs?
# is.na(bike$dateInstalled)
# !is.na(bike$dateInstalled)What is another way to remove zeros from the data?
Add a filtering step
C. What was the average bike lane length grouped by dateInstalled? Remove NAs with na.rm = TRUE.
bike %>%
group_by(dateInstalled) %>%
summarise(mean_of_the_bike = mean(length, na.rm = F))## # A tibble: 9 x 2
## dateInstalled mean_of_the_bike
## <dbl> <dbl>
## 1 2006 1469.
## 2 2007 310.
## 3 2008 249.
## 4 2009 407.
## 5 2010 246.
## 6 2011 233.
## 7 2012 271.
## 8 2013 290.
## 9 NA 215.# Can combine 6 & 7!
mean(bike$length[ !is.na(bike$dateInstalled)])## [1] 273.9943bike %>%
mutate(length = ifelse(length == 0, NA, length)) %>%
group_by(dateInstalled) %>%
summarise(n = n(), # Add this column if you want!
mean_of_the_bike = mean(length, na.rm = TRUE),
n_missing = sum(is.na(length))) # Add this column if you want!## # A tibble: 9 x 4
## dateInstalled n mean_of_the_bike n_missing
## <dbl> <int> <dbl> <int>
## 1 2006 2 1469. 0
## 2 2007 368 310. 0
## 3 2008 206 249. 0
## 4 2009 86 407. 0
## 5 2010 625 246. 0
## 6 2011 101 233. 0
## 7 2012 107 271. 0
## 8 2013 10 290. 0
## 9 NA 126 217. 1D. Does the plot from question 10 improve if you remove the zeros?
plot( pull(bike, dateInstalled), pull(bike, length) )
Yes!
E. What kind of plot would be better for showing the length by each year group? Make this plot.
A boxplot would be more appropriate since year behaves more as a category than a continuous number.
boxplot( pull(bike, length) ~ pull(bike, dateInstalled) )