ASP 460 2.0 Data Visualisation
Calendar-based Time Series Visualization
R Packages
library(tidyverse)
# install.packages("remotes")
# remotes::install_github("earowang/sugrrants")
library(sugrrants)
library(tsibble)
library(lubridate)
theme_set(theme_bw())Data
hourly_peds## # A tibble: 78,755 x 10
## Date_Time Date Year Month Mdate Day Time Sensor_ID
## <dttm> <date> <dbl> <ord> <dbl> <ord> <dbl> <dbl>
## 1 2016-01-01 00:00:00 2016-01-01 2016 January 1 Friday 0 18
## 2 2016-01-01 00:00:00 2016-01-01 2016 January 1 Friday 0 13
## 3 2016-01-01 00:00:00 2016-01-01 2016 January 1 Friday 0 3
## 4 2016-01-01 00:00:00 2016-01-01 2016 January 1 Friday 0 9
## 5 2016-01-01 00:00:00 2016-01-01 2016 January 1 Friday 0 6
## 6 2016-01-01 00:00:00 2016-01-01 2016 January 1 Friday 0 25
## 7 2016-01-01 00:00:00 2016-01-01 2016 January 1 Friday 0 30
## 8 2016-01-01 01:00:00 2016-01-01 2016 January 1 Friday 1 18
## 9 2016-01-01 01:00:00 2016-01-01 2016 January 1 Friday 1 13
## 10 2016-01-01 01:00:00 2016-01-01 2016 January 1 Friday 1 3
## # i 78,745 more rows
## # i 2 more variables: Sensor_Name <chr>, Hourly_Counts <dbl>glimpse(hourly_peds)## Rows: 78,755
## Columns: 10
## $ Date_Time <dttm> 2016-01-01 00:00:00, 2016-01-01 00:00:00, 2016-01-01 00~
## $ Date <date> 2016-01-01, 2016-01-01, 2016-01-01, 2016-01-01, 2016-01~
## $ Year <dbl> 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 20~
## $ Month <ord> January, January, January, January, January, January, Ja~
## $ Mdate <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,~
## $ Day <ord> Friday, Friday, Friday, Friday, Friday, Friday, Friday, ~
## $ Time <dbl> 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2,~
## $ Sensor_ID <dbl> 18, 13, 3, 9, 6, 25, 30, 18, 13, 3, 9, 6, 25, 30, 18, 13~
## $ Sensor_Name <chr> "Collins Place (North)", "Flagstaff Station", "Melbourne~
## $ Hourly_Counts <dbl> 409, 785, 2829, 915, 3643, 2844, 953, 200, 313, 2905, 38~unique(hourly_peds$Sensor_Name)## [1] "Collins Place (North)"
## [2] "Flagstaff Station"
## [3] "Melbourne Central"
## [4] "Southern Cross Station"
## [5] "Flinders Street Station Underpass"
## [6] "Melbourne Convention Exhibition Centre"
## [7] "Lonsdale St (South)"Data Visualization
Plot 1
hourly_peds %>%
filter(Date < as.Date("2016-03-01")) %>%
ggplot(aes(x = Time, y = Hourly_Counts, colour = Sensor_Name)) +
geom_line() +
facet_calendar(~ Date) +
theme_bw() +
theme(legend.position = "bottom")
Plot 2
hourly_peds %>%
ggplot(aes(x = Date, y = Hourly_Counts, colour = Sensor_Name)) +
geom_line(size = 0.3) +
facet_grid(
Sensor_Name ~ .,
labeller = labeller(Sensor_Name = label_wrap_gen(20))
) + theme(legend.position = "bottom") +
xlab("Date Time") +
ylab("Hourly Counts")
Your turn: Level-up plot 2
Plot 3
fs <- hourly_peds %>%
filter(Sensor_Name == "Flagstaff Station")
fs## # A tibble: 11,567 x 10
## Date_Time Date Year Month Mdate Day Time Sensor_ID
## <dttm> <date> <dbl> <ord> <dbl> <ord> <dbl> <dbl>
## 1 2016-01-01 00:00:00 2016-01-01 2016 January 1 Friday 0 13
## 2 2016-01-01 01:00:00 2016-01-01 2016 January 1 Friday 1 13
## 3 2016-01-01 02:00:00 2016-01-01 2016 January 1 Friday 2 13
## 4 2016-01-01 03:00:00 2016-01-01 2016 January 1 Friday 3 13
## 5 2016-01-01 04:00:00 2016-01-01 2016 January 1 Friday 4 13
## 6 2016-01-01 05:00:00 2016-01-01 2016 January 1 Friday 5 13
## 7 2016-01-01 06:00:00 2016-01-01 2016 January 1 Friday 6 13
## 8 2016-01-01 07:00:00 2016-01-01 2016 January 1 Friday 7 13
## 9 2016-01-01 08:00:00 2016-01-01 2016 January 1 Friday 8 13
## 10 2016-01-01 09:00:00 2016-01-01 2016 January 1 Friday 9 13
## # i 11,557 more rows
## # i 2 more variables: Sensor_Name <chr>, Hourly_Counts <dbl>glimpse(fs)## Rows: 11,567
## Columns: 10
## $ Date_Time <dttm> 2016-01-01 00:00:00, 2016-01-01 01:00:00, 2016-01-01 02~
## $ Date <date> 2016-01-01, 2016-01-01, 2016-01-01, 2016-01-01, 2016-01~
## $ Year <dbl> 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 20~
## $ Month <ord> January, January, January, January, January, January, Ja~
## $ Mdate <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,~
## $ Day <ord> Friday, Friday, Friday, Friday, Friday, Friday, Friday, ~
## $ Time <dbl> 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16~
## $ Sensor_ID <dbl> 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, ~
## $ Sensor_Name <chr> "Flagstaff Station", "Flagstaff Station", "Flagstaff Sta~
## $ Hourly_Counts <dbl> 785, 313, 116, 85, 40, 38, 40, 40, 49, 67, 82, 96, 88, 1~fs_cal <- fs %>%
frame_calendar(x = Time, y = Hourly_Counts, date = Date)
fs_cal## # A tibble: 11,567 x 12
## Date_Time Date Year Month Mdate Day Time Sensor_ID
## <dttm> <date> <dbl> <ord> <dbl> <ord> <dbl> <dbl>
## 1 2016-01-01 00:00:00 2016-01-01 2016 January 1 Friday 0 13
## 2 2016-01-01 01:00:00 2016-01-01 2016 January 1 Friday 1 13
## 3 2016-01-01 02:00:00 2016-01-01 2016 January 1 Friday 2 13
## 4 2016-01-01 03:00:00 2016-01-01 2016 January 1 Friday 3 13
## 5 2016-01-01 04:00:00 2016-01-01 2016 January 1 Friday 4 13
## 6 2016-01-01 05:00:00 2016-01-01 2016 January 1 Friday 5 13
## 7 2016-01-01 06:00:00 2016-01-01 2016 January 1 Friday 6 13
## 8 2016-01-01 07:00:00 2016-01-01 2016 January 1 Friday 7 13
## 9 2016-01-01 08:00:00 2016-01-01 2016 January 1 Friday 8 13
## 10 2016-01-01 09:00:00 2016-01-01 2016 January 1 Friday 9 13
## # i 11,557 more rows
## # i 4 more variables: Sensor_Name <chr>, Hourly_Counts <dbl>, .Time <dbl>,
## # .Hourly_Counts <dbl>glimpse(fs_cal)## Rows: 11,567
## Columns: 12
## $ Date_Time <dttm> 2016-01-01 00:00:00, 2016-01-01 01:00:00, 2016-01-01 0~
## $ Date <date> 2016-01-01, 2016-01-01, 2016-01-01, 2016-01-01, 2016-0~
## $ Year <dbl> 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2~
## $ Month <ord> January, January, January, January, January, January, J~
## $ Mdate <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1~
## $ Day <ord> Friday, Friday, Friday, Friday, Friday, Friday, Friday,~
## $ Time <dbl> 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 1~
## $ Sensor_ID <dbl> 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13,~
## $ Sensor_Name <chr> "Flagstaff Station", "Flagstaff Station", "Flagstaff St~
## $ Hourly_Counts <dbl> 785, 313, 116, 85, 40, 38, 40, 40, 49, 67, 82, 96, 88, ~
## $ .Time <dbl> 0.1842593, 0.1857890, 0.1873188, 0.1888486, 0.1903784, ~
## $ .Hourly_Counts <dbl> 0.9556459, 0.9522512, 0.9508343, 0.9506113, 0.9502877, ~p_fs <- fs_cal %>%
ggplot(aes(x = .Time, y = .Hourly_Counts, group = Date)) +
geom_line() +
theme(legend.position = "bottom")
p_fs
prettify(p_fs)
Your turn: Level-up the plot
Plot 4
fs_cal_free <- fs %>%
frame_calendar(x = Time, y = Hourly_Counts, date = Date, scale = "free")
p_fs_free <- fs_cal_free %>%
ggplot(aes(x = .Time, y = .Hourly_Counts, group = Date)) +
geom_line() +
theme(legend.position = "bottom")
prettify(p_fs_free)
Your turn: Level-up the plot
Plot 5
fs_polar <- fs %>%
frame_calendar(x = Time, y = Hourly_Counts, date = Date, polar = TRUE)
p_fs_polar <- fs_polar %>%
ggplot(aes(x = .Time, y = .Hourly_Counts, group = Date)) + geom_path() +
theme(legend.position = "bottom")
prettify(p_fs_polar)
Your turn: Level-up the plot
Plot 6
pedestrian_dec <- hourly_peds %>%
filter(Month == "December") %>%
frame_calendar(
x = Time, y = Hourly_Counts, date = Date, width = 0.97, height = 0.97
)
p_boxplot <- pedestrian_dec %>%
ggplot() +
geom_boxplot(
aes(x = .Time, y = .Hourly_Counts, group = Date_Time),
outlier.size = 0.3, width = 0.004, position = "identity",
colour = "grey50"
)
p_boxplot
prettify(p_boxplot)
Plot 7
p_boxplot_2 <- p_boxplot +
geom_smooth(
aes(.Time, .Hourly_Counts, group = Date),
se = FALSE, method = "loess"
)
prettify(p_boxplot_2)
The content is based on the paper
Wang, E., Cook, D., & Hyndman, R. J. (2020). Calendar-based graphics for visualizing people’s daily schedules. Journal of computational and graphical statistics, 29(3), 490-502.