ETC5512: Wild Caught Data

class: middle center hide-slide-number monash-bg-gray80

.info-box.w-50.bg-white[
These slides are viewed best by Chrome or Firefox and occasionally need to be refreshed if elements did not load properly. See <a href=lecture-05.pdf>here for the PDF </a>. 
]

.white[Press the **right arrow** to progress to the next slide!]

---

class: title-slide
count: false
background-image: url("images/bg-01.png")

# .monash-blue[ETC5512: Wild Caught Data]

<h2 style="font-weight:900!important;">Case Study: Victorian Transport Data</h2>

.bottom_abs.width100[

Lecturer: *Belinda Maher, Department of Transport*

Department of Econometrics and Business Statistics

ETC5512.Clayton-x@monash.edu

Week 5

]

---
#Lecture Contents:

.pull-left[
- About Me

- Open Transport Data Examples (Including GTFS and Shapefiles of the network)

- A little bit about one of my work projects with Google Maps API data

- Melbourne Public Transport

- What is Operational Performance?

- Example of wrangling Public Transport Performance Data from an Excel file (using the tidyverse)

- Plotting this data with ggplot2
]
.pull-right[<a title="JohnnoShadbolt at English Wikipedia, CC BY 3.0 &lt;https://creativecommons.org/licenses/by/3.0&gt;, via Wikimedia Commons" href="https://commons.wikimedia.org/wiki/File:Melbourne_trams_map_old.gif"><img width="512" alt="Melbourne trams map old" src="https://upload.wikimedia.org/wikipedia/commons/thumb/4/4a/Melbourne_trams_map_old.gif/512px-Melbourne_trams_map_old.gif"></a>]
---
.grid[
.item[
#About Me

####Belinda Maher

I ❤️

Data, R, Maps & Trains
 
Background: 
 
- KiwiRail (Wellington) (2009-2012)
 - Auckland Electrification & Track Upgrades 
- Grad Dip Statistics & Operations Research
- Department of Transport (2016 - Current) 
 - Operational Performance Analysis (PTV)
 - Strategic Transport Network Planning
 - Bus & Tram Planning
]
.item.right[
<img src="images/lecture-05/AT EMU s.png"/>
]
]

???
image source: https://en.wikipedia.org/wiki/New_Zealand_AM_class_electric_multiple_unit#/media/File:AMA_103_at_Puhinui.jpg

---
class:transition
#Open Transport Data
##Example: What can you achieve with open data and a bit of passion?

---

#Is the 370 the worst bus route in Sydney?

Excellent example of wild caught data

.pull-left[<a href="https://youtu.be/O7jqU39wvKk"><img src="images/lecture-05/Katie Bell.jpg" width=100%/> </a>
]
.pull-right[
- Katie used 130GB of GTFS-Realtime data
- She analysed bus lateness in Sydney over 1 year
 (Detailed Sydney lateness data isn't reported publicly) 
[YouTube Link](https://youtu.be/O7jqU39wvKk) 
[Yow Conference Slides](https://yowconference.com/talks/katie-bell/yow-2018-sydney/is-the-370-the-worst-bus-in-sydney-8530/)
&nbsp; https://github.com/katharosada/bus-shaming
]
---

class:transition
## My biggest R project
# Google Maps API data (predicted travel times) 
##(not open data, but can be obtained for free)

---

class: no-logo
background-position: center top;
background-size:100%
background-image: url("images/lecture-05/Destination 1 Spring St Fullscreen.png")

#library(googleway) and Driving vs Public Transport Travel Times

<div style="position: absolute; bottom: 5px;">
<a href="http://belinda-maher-melburn.netlify.app">Link: Presentation - Visualising predicted travel time differences between transport modes </a>
</div>
---
class:transition

#Geospatial & Timetable Data

---
#General Transit Feed Specification (GTFS)

GTFS contains timetable, stop and path data for PT services in Victoria (and is a global standard)
 
It's the data which Google Maps ingests to show you PT in Google Maps

I go through using GTFS shape and stop data in R in an RLadies talk I did in 2018: 
[Slides - Public Transport Maps & Geospatial Data in R](https://drive.google.com/file/d/191Jq1awInmtXPXrVlywXXaQjxv0basHe)

There are packages like [tidytransit](http://tidytransit.r-transit.org/index.html) which can help you navigate the GTFS Format (it's a bit [complicated](https://xang1234.github.io/images/isochrone/gtfs.png))

For example you could look into service frequencies & headways for each public transport route - 
http://tidytransit.r-transit.org/articles/frequency.html

---
class: no-logo
background-position: center top;
background-size:100%
background-image: url("images/lecture-05/Melbourne.JPG")

#The Melbourne Public Transport Network

---

# PTV shapefiles - paths and points in a file for GIS
All the shapes in previous map can be found at: 
https://www.ptv.vic.gov.au/footer/data-and-reporting/datasets/

You can also get all the stop locations easily out of GTFS text files without using shapefiles

Of the many shapefiles available, you might be interested in these: 
.grid[
 .item[
Paths: 
<a href="https://discover.data.vic.gov.au/dataset/ptv-train-corridor-centreline">Train Corridor Centreline</a> 
<a href="https://discover.data.vic.gov.au/dataset/ptv-metro-tram-routes">Metro Tram Routes</a> 
<a href="https://discover.data.vic.gov.au/dataset/ptv-metro-bus-routes">Metro Bus Routes</a> ]

.item[
 Stops/Stations as shapefiles: 
<a href="https://discover.data.vic.gov.au/dataset/ptv-metro-train-stations">Metro Train Stations</a> 
<a href="https://discover.data.vic.gov.au/dataset/ptv-metro-tram-stops">Metro Tram Stops</a> 
<a href="https://discover.data.vic.gov.au/dataset/ptv-metro-bus-stops">Metro Bus Stops</a> 
]]

???
from https://belinda-maher-melburn.netlify.app/#41
---
class:transition
#Other Available Transport Data
##(quick mentions)
---
# Further Relevant Data Links

Metro Bus Performance data has been requested by Philip Mallis under Freedom of Information:
https://philipmallis.com/blog/2020/02/28/melbourne-bus-route-patronage-data-for-2008-2019/

Public Transport Patronage: 
[Patronage data via Philip Mallis (some is also available from PTV)](
https://philipmallis.com/blog/2019/02/14/station-patronage-in-victoria-2013-2018/)

[Patronage from PTV](https://transport.vic.gov.au/about/data-and-research/patronage)

Roads: 
[Road Network & Performance (VicRoads)](https://www.vicroads.vic.gov.au/traffic-and-road-use/road-network-and-performance)

- [Including Road Use & Performance](https://www.vicroads.vic.gov.au/traffic-and-road-use/road-network-and-performance/road-use-and-performance)

If you wanted to look further at the effects of COVID restrictions:
 
COVID Restriction Timeline (also used by Melbourne on Transit)
https://deborahalupton.medium.com/timeline-of-covid-19-in-australia-1f7df6ca5f23

---
class:transition

#(Public) Public Transport Data

---

#Public Transport Modes
<div>
We're going to look at the rail modes:
.grid[.item[<center>Metro Trains 🚆 <img src="images/lecture-05/Metro.png"/></center>]
.item[<center>Yarra Trams 🚋 <img src="images/lecture-05/Yarra Trams.png"/></center>]
.item[<center>V/Line Trains 🚂 <img src="images/lecture-05/VLine.jpg" width=40%></center>]]
 
Performance data for Bus & Coach is not easily available from the PTV Website PowerBI app:

.grid[
.item[Metro Bus & Regional Bus <img src="images/lecture-05/PTV Bus.jpg" width=100%/>]
.item[V/Line Coach <img src="images/lecture-05/Wong VLine Coach.jpg" width=100%/>]]
</div>

---
#Public Transport Operational Performance
So, what is operational performance?

.pull-left[
.grid[
 .item[
 <img src="images/lecture-05/Tram OP.jpg"/>
 ]
 .item[
 <img src="images/lecture-05/Train OP.jpg" width=100%/>
 ]]]
 .pull-right[<img src="images/lecture-05/YT Perf.jpg"/ width=100%>]
 
 .footnote[.f5[graph source: https://yarratrams.com.au/performance-results]]

---
#Punctuality and Reliability

Official Definition
.blockquote[

<h2>What do we monitor?</h2>Every operator needs to meet the performance thresholds in their contact. The thresholds are:<ul><li>Punctuality - the percentage of on-time events at specific points.</li>
<li>Reliability - the proportion of timetabled services the operator delivers.</li>
</ul>.cite[PTV Website &nbsp; https://www.ptv.vic.gov.au/footer/data-and-reporting/network-performance/]]

![PTV](images/lecture-05/PTV.jpg)
Public Transport Victoria is part of the Victorian Government Department of Transport and manages PT service delivery through franchises with operators
---

#Unpacking the Measurement of Public Transport Services

.grid[
.item[
##Punctuality:

Adherence to timetable

When? 
Compare:

Timetabled Arrival / Departure Time 
to Actual Arrival /
Departure Time

Where? 
At timing points or destination
 
].item[
##Reliability:
Did the service run? 
- Cancelled Services

- 'Short' Services

- Diverted Services

- Loop Skips: 
 City Loop Westona Loop
].item[<img src="images/lecture-05/rolling stock.jpg">]
]

.footnote[[Image source](https://www.ptv.vic.gov.au/assets/default-site/more/travelling-on-the-network/accessibility/Try-Before-You-Ride/7d5561782e/Try-before-you-ride-header.png)]

???

Why do we have to include Loop Skips & Altered Expresses
---
#Monthly Performance Reports on PTV Website

.footnote[http://www.ptv.vic.gov.au/footer/data-and-reporting/network-performance/monthly-performance/]
---
#Daily Performance Reports on PTV Website

.grid[.item[ Daily figures are available <a href="https://www.ptv.vic.gov.au/footer/data-and-reporting/network-performance/daily-performance/">here</a>

- These daily figures could be scraped and collected weekly for more detailed data analysis than monthly figures.

- We can't get line level information from the website but it is available in a PowerBI App
]
.item[<img src="images/lecture-05/Daily OP.jpg" width=90%/> ]]

---
class: monash-bg-blue2 white
#PTV PowerBI app for Performance Data
The link to the dashboard is located at the bottom of the .monash-black.white[<a href="https://www.ptv.vic.gov.au/footer/data-and-reporting/network-performance/monthly-performance/">monthly performance page</a>]
<img src="images/lecture-05/Monthly OP Cover.JPG" width=70%/>
.footnote[<a href="https://app.powerbi.com/view?r=eyJrIjoiMzE2ZDIyN2YtODY1Yy00ZGY0LWE4YTktNDcxOTcwYWQyMjM5IiwidCI6IjcyMmVhMGJlLTNlMWMtNGIxMS1hZDZmLTk0MDFkNjg1NmUyNCJ9"&gt;PTV PowerBI App Link</a>]

---
class: center
<img src="images/lecture-05/Tram PowerBI.JPG" width=90%>

---
class:transition

#Deriving meaning from the data
---
#Melbourne Transport Bloggers
.pull-left[
Chris Loader ([Charting Transport](https://chartingtransport.com/)) and Peter Parker ([Melbourne On Transit](https://melbourneontransit.blogspot.com/))

[Charting Transport COVID and pedestrian volumes](https://chartingtransport.com/2020/05/25/what-impact-has-the-2020-covid-19-pandemic-had-on-pedestrian-volumes-in-central-melbourne/)

(Di has already mentioned the [City of Melbourne Pedestrian Counting System](http://www.pedestrian.melbourne.vic.gov.au/?_ga=2.225968922.1626262524.1612686930-400371523.1612686930#date=07-02-2021&time=18)), which Charting Transport has analysed in the above post
 
This data could potentially be used to give a baseline level of people in the city during COVID restrictions if you wanted to compare it to other data
]
.pull-right[<img src="images/lecture-05/CL pedestrian.PNG" width=100%>]
.footnote[graph source: Chris Loader - http://chartingtransport.com ]
---

####Melbourne Metro Train Performance
 We're going to talk tonight about some of the data related to this blog post: 
[Melbourne On Transit: How are our trains performing with hardly any passengers?](https://melbourneontransit.blogspot.com/2021/02/timetable-tuesday-105-how-do-our-trains.html)

.footnote[Image, regression and callout box source:&nbsp;<a href="https://1.bp.blogspot.com/-CUJ4RNYZ9ck/YCmeAQi8pSI/AAAAAAAAHLM/JaxS0vLIbQk9pVkdm1m1hnS3H1j53IK5ACLcBGAsYHQ/s1386/long%2Bterm%2Bperformance.PNG">Melbourne on Transit</a>, not the views of the lecturer!]

???

Think about the regression line

---
class: transition
###Causes of Variation in Performance
.question-box.w-70[What sort of variables might affect the performance of public transport? ]

Please put some ideas in the chat (inventive answers are welcome!)

---
class: monash-bg-green2 white
#Causes of Disruption to Services
####Some of the many reasons public transport services are more late or cancelled than usual:

Traffic, 
&nbsp;Overcrowding,
&nbsp;&nbsp;Suspension of Other Services, 
&nbsp;&nbsp;Rain (and Flooding),&nbsp;&nbsp;
Roadworks,&nbsp;&nbsp; 
Trespassers,&nbsp;&nbsp; 
Suicides,&nbsp;&nbsp; 
Police Requests,&nbsp;&nbsp; 
Hot Weather,&nbsp;&nbsp; 
Temporary Speed Restrictions (Track Condition),&nbsp;&nbsp;
Signal Faults,&nbsp;&nbsp;
Track Faults,&nbsp;&nbsp;
Trees Falling on Track/Overhead,&nbsp;&nbsp;
Overhead Faults,&nbsp;&nbsp; 
Vehicle Faults,&nbsp;&nbsp; 
Power Faults,&nbsp;&nbsp; 
Signals Passed at Danger,&nbsp;&nbsp; 
Grass Fires,&nbsp;&nbsp; 
Driver Shortages,&nbsp;&nbsp; 
Protests,&nbsp;&nbsp; 
Cars Blocking Track,&nbsp;&nbsp; 
Derailments,&nbsp;&nbsp; 
Collisions,&nbsp;&nbsp; 
Evacuation of Train Control,&nbsp;&nbsp; 
Rats Eating Signalling Cables,&nbsp;&nbsp;
A Bat Being Electrocuted By Signalling Equipment,&nbsp;&nbsp;
Tumbleweeds (Fairy Grass),&nbsp;&nbsp;
Millipedes.

####Things which might improve PT performance: (relative to a 'normal' day)
Weekends,&nbsp;&nbsp; Public Holidays,&nbsp;&nbsp; School Holidays,&nbsp;&nbsp; University Holidays,&nbsp;&nbsp; January,&nbsp;&nbsp; Rostered Days Off,&nbsp;&nbsp; COVID-19 Restrictions,&nbsp;&nbsp; Working From Home,&nbsp;&nbsp; Day before Cup Day,&nbsp;&nbsp; etc.
 
####Seasonality of crowding and traffic -> Seasonality of baseline performance
---

#Did you just say Millipedes?

https://www.theage.com.au/national/victoria/trekking-millipedes-sideline-trains-20120328-1vyri.html

and...[a link to the poor bat](https://www.theage.com.au/national/victoria/train-chaos-as-faults-cause-long-delays-20130205-2dva2.html)

"Suffice to say, the bat is no longer with us. It was his untimely 
departure that has caused the extent of the damage."

---

#Munging Network Performance Data

So, let's have a look at some performance data from PTV from that PowerBI App.

Downloading is disabled on the app, but an Excel file will be provided.

Why Excel? Because you need to learn how to deal with data which is not in an ideal format :)

We're going to need to use some R packages to manipulate this data...

###Tools of the Trade: The tidyverse (particularly dplyr, magrittr & ggplot2) 
.grid[
.item[
<img src="images/lecture-05/hex-tidyverse.png" width=70%/>]
.item[<img src="images/lecture-05/dplyr.PNG" width=70%/>]
.item[<img src="images/lecture-05/pipe.PNG" width=70%/>]
.item[<img src="images/lecture-05/ggplot2.PNG" width=70%/>]

]
.footnote[https://www.tidyverse.org/packages/ 
https://magrittr.tidyverse.org/articles/magrittr.html
]

---
#loading data in to R from Excel 
####using readxl::read_excel()

```r
library(here)
library(readxl)
library(curl)
#Download and save the file from
url<-"https://wcd.numbat.space/lectures/data/lecture-05/Rail%20PT%20Public%20Monthly%20Figures%20cleaned.xlsx"
curl_download(url, "Rail PT Public Monthly Figures cleaned.xlsx")
```

```r
#dir(here('lectures','data','lecture-05'))
filename<-here('lectures','data','lecture-05',#folders for knitting lecture
 'Rail PT Public Monthly Figures cleaned.xlsx')
sheets<-excel_sheets(filename)
#This gives us the names of the sheets within this Excel file
sheets
```

```
## [1] "Metro" "Yarra" "VLine"
```

---

#Reading in the data with library(purrr)

```r
#read all three sheets in to df
library(purrr)
performance <- map(sheets, ~read_xlsx(filename, sheet=.x))

#The above line does exactly what these 3 lines do, 
#but automatically for every sheet in the Excel workbook.
 performance2 <- list(read_xlsx(filename, sheet="Metro"),
 read_xlsx(filename, sheet="Yarra"),
 read_xlsx(filename, sheet="VLine"))

#compare::compare(performance,performance2)
## TRUE
#Name the items in the list (created from the tabs in the spreadsheet)
names(performance)<-sheets
names(performance)
```

```
## [1] "Metro" "Yarra" "VLine"
```

---
#bind_rows() - join the data into one data frame

```r
library(dplyr)
performance <- bind_rows(performance, .id="Mode")
performance <- select(performance,all_of(keep))

head(performance)
```

```
## # A tibble: 6 x 15
## Parent_Route Mode Service_Year Service_Date Service_Month
## <chr> <chr> <dbl> <dttm> <dbl>
## 1 Alamein Metro 2010 2010-01-01 00:00:00 1
## 2 Belgrave Metro 2010 2010-01-01 00:00:00 1
## 3 Craigieburn Metro 2010 2010-01-01 00:00:00 1
## 4 Cranbourne Metro 2010 2010-01-01 00:00:00 1
## 5 Frankston Metro 2010 2010-01-01 00:00:00 1
## 6 Glen Waverley Metro 2010 2010-01-01 00:00:00 1
## # ... with 10 more variables: Reporting_Group <chr>,
## # num_services_scheduled <dbl>, num_services_cancelled <dbl>,
## # num_services_short <dbl>, num_services_bypassed <dbl>,
## # num_services_scheduled_at_dest <dbl>, num_services_on_time_at_dest <dbl>,
## # num_observations_on_time <dbl>, num_observations_scheduled <dbl>,
## # pct_on_time_avg <dbl>
```

---

#Summarising the data by year - group_by() and summarise()

```r
library(dplyr)
library(magrittr)
performance %>% 
  #dplyr::filter(Service_Year>2015) %>% 
  group_by(Service_Year, Mode) %>% 
summarise(num_services_scheduled=sum(num_services_scheduled)) %>% 
  ungroup -> perf_year

head(perf_year)
```

```
## # A tibble: 6 x 3
## Service_Year Mode num_services_scheduled
## <dbl> <chr> <dbl>
## 1 2010 Metro 674189
## 2 2010 VLine 74227
## 3 2010 Yarra 1619282
## 4 2011 Metro 706765
## 5 2011 VLine 73916
## 6 2011 Yarra 1637585
```

---
#What variables and routes do we have?
.f4[

```r
names(performance)
```

```
##  [1] "Parent_Route"                   "Mode"                          
##  [3] "Service_Year"                   "Service_Date"                  
##  [5] "Service_Month"                  "Reporting_Group"               
##  [7] "num_services_scheduled"         "num_services_cancelled"        
##  [9] "num_services_short"             "num_services_bypassed"         
## [11] "num_services_scheduled_at_dest" "num_services_on_time_at_dest"  
## [13] "num_observations_on_time"       "num_observations_scheduled"    
## [15] "pct_on_time_avg"
```

```r
#How many routes are there by mode?
with(unique(performance %>% select(Mode, Parent_Route)), table(Mode))
```

```
## Mode
## Metro VLine Yarra 
## 16 11 26
```
]
---
#26 Lines are not easy to read! - Maybe we need to summarise all routes
.f5[
<img src="images/lecture-05/tram_punc-1.png" width="100%" style="display: block; margin: auto;" />
]
---
#Summarising multiple columns
.f4[

```r
library(tidyverse)
library(magrittr)
perf_month <- performance %>% 
 group_by(Service_Date, Service_Year, Service_Month, Mode) %>% 
 summarise(across(starts_with("num_"), sum, .names = "{.col}")) %>%
ungroup()
head(perf_month)
```

```
## # A tibble: 6 x 12
## Service_Date Service_Year Service_Month Mode num_services_scheduled
## <dttm> <dbl> <dbl> <chr> <dbl>
## 1 2010-01-01 00:00:00 2010 1 Metro 56500
## 2 2010-01-01 00:00:00 2010 1 VLine 6167
## 3 2010-01-01 00:00:00 2010 1 Yarra 134467
## 4 2010-02-01 00:00:00 2010 2 Metro 51598
## 5 2010-02-01 00:00:00 2010 2 VLine 5672
## 6 2010-02-01 00:00:00 2010 2 Yarra 125090
## # ... with 7 more variables: num_services_cancelled <dbl>,
## # num_services_short <dbl>, num_services_bypassed <dbl>,
## # num_services_scheduled_at_dest <dbl>, num_services_on_time_at_dest <dbl>,
## # num_observations_on_time <dbl>, num_observations_scheduled <dbl>
```
]
---

#How do we re-calculate the summary measures at different grouping levels?
See glossary tab in the dashboard. (It's complicated, so I've done it for you on the next slide!)
<img src="images/lecture-05/Glossary Metro.JPG" width=90% />

---

#Punctuality and Reliability calculations in R
(don't worry about this detail, I've written the function for you)
.f5[

```r
perf_calc<- perf_month %>% mutate(
 Reliability=case_when(
 Mode =="Metro" ~ ((1-((num_services_cancelled+(0.25*num_services_short)+(0.125*num_services_bypassed))
 /num_services_scheduled))*100),
 Mode == "Yarra" ~ (1-((num_services_cancelled+ 0.125*num_services_short)
 /num_services_scheduled))*100,
 Mode == "VLine" ~ (1-(num_services_cancelled/num_services_scheduled))*100),

Punctuality=case_when(
  Mode =="Metro" ~ (num_services_on_time_at_dest/num_services_scheduled_at_dest)*100,
  Mode == "Yarra" ~ (num_observations_on_time / num_observations_scheduled)*100,
  Mode == "VLine" ~  (num_services_on_time_at_dest/num_services_scheduled_at_dest)*100
  )
)
```
]
---

#Next step: Tidying the Data
.f4[

```r
library(dplyr)
perf_selected <- perf_calc%>% select("Service_Date","Service_Year","Service_Month",
 "Mode","Reliability", "Punctuality")

perf_selected
```

```
## # A tibble: 400 x 6
## Service_Date Service_Year Service_Month Mode Reliability Punctuality
## <dttm> <dbl> <dbl> <chr> <dbl> <dbl>
## 1 2010-01-01 00:00:00 2010 1 Metro 98.3 86.9
## 2 2010-01-01 00:00:00 2010 1 VLine 98.4 83.1
## 3 2010-01-01 00:00:00 2010 1 Yarra 99.5 84.9
## 4 2010-02-01 00:00:00 2010 2 Metro 98.8 83.4
## 5 2010-02-01 00:00:00 2010 2 VLine 97.5 82.1
## 6 2010-02-01 00:00:00 2010 2 Yarra 99.0 80.8
## 7 2010-03-01 00:00:00 2010 3 Metro 98.6 80.5
## 8 2010-03-01 00:00:00 2010 3 VLine 98.1 79.5
## 9 2010-03-01 00:00:00 2010 3 Yarra 98.8 79.5
## 10 2010-04-01 00:00:00 2010 4 Metro 99.4 84.5
## # ... with 390 more rows
```
]
---

#Tidy Data - pivot_longer()

```r
library(tidyr)
perf_tidy<-perf_selected %>% 
 pivot_longer(cols=c(Reliability, Punctuality),names_to="Measure")
#round the observations to 1 decimal place
perf_tidy$value %<>% round(1)

head(perf_tidy)
```

```
## # A tibble: 6 x 6
## Service_Date Service_Year Service_Month Mode Measure value
## <dttm> <dbl> <dbl> <chr> <chr> <dbl>
## 1 2010-01-01 00:00:00 2010 1 Metro Reliability 98.3
## 2 2010-01-01 00:00:00 2010 1 Metro Punctuality 86.9
## 3 2010-01-01 00:00:00 2010 1 VLine Reliability 98.4
## 4 2010-01-01 00:00:00 2010 1 VLine Punctuality 83.1
## 5 2010-01-01 00:00:00 2010 1 Yarra Reliability 99.5
## 6 2010-01-01 00:00:00 2010 1 Yarra Punctuality 84.9
```

---

#ggplot2 Reliability by Year
.f4[
.pull-left[

```r
library(ggplot2)

reldata <- dplyr::filter(perf_tidy,
 Mode=="Metro",
 Service_Year >= 2016, 
 Measure=="Reliability")

ggplot(data=reldata,
  aes(x=factor(
    month.abb[Service_Month],
    levels=month.abb), 
  y=value, 
  colour=as.character(Service_Year),
  group=Service_Year))+ labs(
    title="Metro Trains Reliability",
    x="Month", colour="Year", 
    y="Reliability") +
  geom_line() + 
  theme(legend.position="bottom")
```
]] .pull-right[
![](images/lecture-05/relplot-1.png)
]

---
#ggplot2 Punctuality by Year
.pull-left[
.f4[

```r
ggplot(data=dplyr::filter(perf_tidy,
  Mode=="Metro", 
  Service_Year >= 2016, 
  Measure=="Punctuality"),
  aes(x=factor(month.abb[Service_Month],
    levels=month.abb), 
    y=value, 
    colour=as.character(Service_Year), 
    group=Service_Year))+
  labs(title="Metro Trains Punctuality",
       x="Month", colour="Year", y="Punctuality") + 
  geom_line()+
  theme(legend.position="bottom")
```
]
]
.pull-right[
![](images/lecture-05/punplot-1.png)
]
---
####Comparing between modes 
.f5[

```r
ggplot(dplyr::filter(perf_calc, Service_Year >= 2017 & Service_Year <= 2020), aes(x=Service_Date, y=Punctuality, 
 colour=Mode, group= Mode)) + geom_line() + labs( title="Punctuality by Mode", x="Month", y="Punctuality")
```

<img src="images/lecture-05/mode-1.png" width="100%" style="display: block; margin: auto;" />
]
---
#facet_wrap()
.f5[

```r
ggplot(dplyr::filter(perf_tidy,Service_Year >= 2018 & Service_Year <= 2020),
 aes(x=factor(month.abb[Service_Month],levels=month.abb), y=value, colour=as.character(Service_Year), 
 group= paste(Mode, Measure, Service_Year)))+geom_line() + 
 facet_wrap(~Mode) + theme(legend.position="bottom") + labs(x="Month", y="Performance", colour="Year") 
```

<img src="images/lecture-05/facet_wrap-1.png" width="100%" style="display: block; margin: auto;" />
]
---

#facet_grid(y~x)

```r
dplyr::filter(perf_tidy, Service_Year >= 2016 & Service_Year <= 2020) %>% 
ggplot(aes(x=factor(month.abb[Service_Month],levels=month.abb), 
 y=value, 
 colour=as.character(Service_Year),
 group= paste(Mode, Measure, Service_Year))) +
 geom_line() + 
 facet_grid(Measure~Mode)+
 theme(legend.position="bottom")+
 labs(x="Month", y="Performance", group="Measure", colour="Year")
```

---
#facet_grid(y~x)

---
class: center, middle
background-image: url("images/lecture-05/FSS.jpg")
background-size: contain
class: monash-bg-black white
&nbsp; &nbsp; &nbsp; &nbsp; &nbsp;  &nbsp; &nbsp; &nbsp; &nbsp; .f2[Questions?]

---

background-size: cover
class: title-slide
background-image: url("images/bg-03.png")

<a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/"><img alt="Creative Commons License" style="border-width:0" src="https://i.creativecommons.org/l/by-sa/4.0/88x31.png" /></a> This work is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/">Creative Commons Attribution-ShareAlike 4.0 International License</a>.

.bottom_abs.width100[

Lecturer: *Belinda Maher, Department of Transport*

Department of Econometrics and Business Statistics

ETC5512.Clayton-x@monash.edu

Week 5

]