RSMG 10 October 2019 ORR protects the interests of rail and road - - PowerPoint PPT Presentation
ORR protects the interests of rail and road users, improving the safety, value and performance of railways and roads today and in the future RSMG 10 October 2019 ORR protects the interests of rail and road users, improving the safety, value and
ORR protects the interests of rail and road users, improving the safety, value and performance of railways and roads today and in the future
ORR protects the interests of rail and road users, improving the safety, value and performance of railways and roads today and in the future
Reproducible Analytical Pipelines Lucy Charlton
3
4
5
6
7
8
9
10
11
ORR protects the interests of rail and road users, improving the safety, value and performance of railways and roads today and in the future
Greg Williams
13
The focus of this presentation is on Web Scraping, but the automation of the production of the Fares Index is included as an example of an alternative application of Python in automating existing process. This work is part of civil service efforts to drive efficiency gains by producing Reproducible Analytical Pipelines. (RAP) https://gss.civilservice.gov.uk/blog/learning-about-reproducible- analytical-pipelines-rap-two-weeks-with-the-gss-good-practice-team/ Why automate at all?
14
What’s involved
individual steps to follow
15
Red items are outputs not produced in manual process:
16
Web scraping, web harvesting, or web data extraction is data scraping used for extracting data from websites.[1] Web scraping software may access the World Wide Web directly using the Hypertext Transfer Protocol, or through a web browser. While web scraping can be done manually by a software user, the term typically refers to automated processes implemented using a bot or web crawler. It is a form of copying, in which specific data is gathered and copied from the web, typically into a central local database or spreadsheet, for later retrieval or analysis.
Source: https://en.wikipedia.org/wiki/Web_scraping
This presentation will cover web scraping through http, rather than an API Direct webscraping via a URL: https://www.nationalrail.co.uk/
17
Economists want to see if the introduction of Open Access Operators changes availability of tickets and/or prices of tickets. This means collecting individual ticket prices based on particular routes, days of the week and times of day. Manual Data Collection: go into the NRE website, request data for route time combination and copy down the results into a spreadsheet.
18
Automatic Data Collection through web scraping:
being automatically appended
and searches for appropriate times for each route
without human intervention – no transcription errors
be made
weekend data is captured
19
This has been rolled out to 6 ORR economists Meeting with DfT Data Science team for a technical demonstration/explanation on 16th October
changes in tickets a given number days from the present day changes in tickets for fixed day in the future over a time series
20
We are ‘the glue that holds the network together’, enabling the seamless provision of cross-boundary services and coordinating capacity requests to make the best use of the network. Our vision is to become the recognised expert trusted by decision makers to plan Britain’s railway. To achieve this we provide a wholesystem, long term view, using the detailed knowledge we have from capacity planning and timetabling the network. Our work is informed by our industry-wide interfaces with every train operator, the Network Rail routes, other infrastructure managers, public and private funders, and franchising bodies.
Idea; The proposal is to work collaboratively with RDG, TfL and others to develop a Digital twin of railway operations; including nowcasting, forecasting and scenario simulation capabilities. The resulting digital asset will provide a Rail Platform for Mobility as a Service (MAAS) that maximises benefits to all travellers and enables more optimal management of the railway including the avoidance or delay of enhancements through operational interventions and safer and more effective management of perturbation. It will include the collection of passenger journey data that enables creation of new KPIs that more closely reflect passenger experience. The work will be aligned with digital twin activities being undertaken by Network Rail
Initial Stakeholders: Supported by Phani Chinchapatnam (Enterprise Architect) and Zach Naylor (Principal Engineer Systems Engineering). David Harding - Head of Economics and Analysis Stephen Draper - Performance Analysis Manager NR Franchise Team TfL DfT RDG
itial Requirements from Economics and Analysis Team; In
Funding Request; £50k Outputs;
all expected use cases 2. Initial Conversations with DISIC 3. Creation of high level system requirements & review with stakeholders 4. First characterisation of benefits against use cases 5. Benefit approximation using PDFH and Green Book 6. Consideration of internal and external funding options including potential for cost sharing by collaboration partners 7. Develop and R&D project plan consistent with proposed funding approach
ORR protects the interests of rail and road users, improving the safety, value and performance of railways and roads today and in the future
29
30
Every year we conduct around 60,000 interviews with passengers – the National Rail Passenger Survey (NRPS)
recognition platform – to identify themes and insights that we could all learn from
can be mined for new insight, ongoing
ratings system – they reflect the broader contextual things that are on rail passengers’ minds…
31
32
The comments are not entirely representative or balanced: people with negative experiences are more likely to add a comment. However, this is important because it helps us learn the ‘why’. Hence – qualitative, but at scale
People tend to use the opened ended question to either reinforce points made, or highlight contextual issues
51 54 64 77 82 10 20 30 40 50 60 70 80 90 Very Satisfied Fairly Satisfied Neither/nor Fairly Dissatisfied Very Dissatisfied
% giving open comment Answer at Q16: Satisfaction with journey
Likelihood of meaningful comment
N = 34,720 40,967 9,002 4,703 2,245
Total comments 51,674 across last four waves
New generation AI platform to derive meaning and themes from unstructured open text data
Signal from Noise Revelation, not search Storytelling & interrogation
Accelerated reading; fast, rigorous decoding of messy data revealing: meanings, emotional energies, attitudes and feelings; complex energy rather than base sentiment; meaning rather than simple coding Surfacing naturally emerging patterns and themes using neural nets, machine learning models, and advanced analytics. Shows what is there rather than just finding things you need to tell it to look for Cutting analysis time and cost by removing a lot of the grunt work – allowing human minds to do more of what they’re best at – thinking and interpretation
3 3
34
Everything’s Fine!
Signoi then uses a form of cluster analysis to automatically identify and quantify themes. Human analysts then name the themes based on the nature of the comments
Interconnected Frequency Dirty trains No Complaints Basic facilities Old & Shabby Crush Hour Misery Generally unreliable No hassle Accessibility and space Travelling (dis)comfort Basic Efficiency Human touch Feeling safe Misleading information Peace & quiet Mobile office Seating management Atypicality Satisfied BUT Signaling issues Value for money Announcements Faster trains Fares What’s going on? Transport options General annoyance Renationalise! Ticketing Minor quibbles Very expensive
8.8 7.8 10
Topic Presence %
7.2 6.9 5.3 4.9 4.4 4.1 3.9 3.9 3.7 3.2 2.9 2.9 2.8 2.6 2.3 2.1 2.1 2.0 2.0 1.9 1.9 1.8 1.7 1.5 1.4 1.1 0.8 0.7 0.6 0.5 0.3 2 4 6 8
NOTE: Online dashboards allow you to explore the themes in detail, by TOC, journey type, and so on
35
TOPIC IN A NUTSHELL
frequency Issues with short connection times or lack of frequent connections, delays have knock on effect
Trains are not clean, sometimes smell, often old
Literally, no complaints!
On train and in station, availability of toilets, catering etc
Rolling stock is tired, rickety, and uncomfortable
Debilitating delays and cancellations, often frequent
Minor issues that do not impact satisfaction
Often contextualising against small delays
Explicit mentions of signal problems, track trespass, etc
Bland positive assessments
space Complex topic containing accessibility in all its forms: luggage, buggies, bikes, car parking, even escalators
discomfort Uncomfortable train environment – seats, temperature, etc
efficiency Getting from A to B as advertised
Complaints about ticket prices
Stories of hellish rush hour journeys
Mentions of staff (station & train) usually, not always, positive
unreliable Wider contextual comments about a route or an
TOPIC IN A NUTSHELL
Making point that while this journey may have been OK, it is generally worse
management Availability of seat bookings, and whether they are honoured on the train itself
Positive comments about a pleasant journey
Generally negative about value for money relative to service levels delivered
Unclear/inaudible announcements (train/station)
General comments about slowness on some routes
Complexity and/or price of getting the right fare
Confusing or absent information (especially stations) when trains are delayed / cancelled
Concerns about safety on board and at stations, rowdy passengers, need for guards often mentioned
Lack of choice in how to get from A to B
annoyance A variety of irritations that do not sit well within other topics
information A major cause of stress, especially at stations – wrong information about platforms, etc
Belief that TOCs need for profit undermines service
The need for quiet journeys – includes issues with noisy passengers, quiet zones, etc
Mainly about machines and lack of clarity about what ticket is valid on which route
Onboard WiFi, power, etc, mainly for business people
NOTE: The themes aggregate comments together based on tone of voice as much as detailed content
36
EXAMPLES
NOTE: Online dashboards allow you to explore the themes in detail, by TOC, journey type, and so on
General unreliability is a strong driver of (lack of) trust
Southern Rail provide a terrible service and the franchise should be awarded to a train company who can provide a decent service The rail service is awful, commute regularly and cannot rely on the trains to be on time, terrible Generally, trains running through Streatham are late or cancelled at very short notice. A very unreliable rail service This train (the 16.44) isn't too bad, but the DRS operated trains have terrible reliability and are slower Looking forward to Crossrail - hope its an improvement. Heathrow Connect frequently cancels services The rail service is appalling and overpriced for the service received Govia provide a terrible service, overpriced. I have changed how I live my life in order to get to work! West Croydon is fine. Southern trains are extremely unreliable though and the service is generally shambolic Thameslink is poor and unreliable, constantly late and poor quality trains Arriva Trains is very unreliable and makes no compensation. Overpriced for such poor service
37
EXAMPLES
NOTE: Online dashboards allow you to explore the themes in detail, by TOC, journey type, and so on
The human touch is very important to passengers
One of the staff at Sittingbourne helped another passenger with a young child on to the train and asked if she needed assistance at the other end. He was very helpful to her and it didn't go unnoticed The staff on the train are always polite and friendly and very informative, always helpful As a solo traveler I always appreciate being able to see a member of staff on the train The lady member of staff at the station was extremely helpful and polite Staff were extremely helpful and polite on the train and on the platform/station Train and train station staff are always pleasant Friendly staff at Beaconsfield station - always helpful and kind Station staff very helpful especially two ladies on Clapham station
38
In terms of tangible themes, commuters, business travellers and leisure travellers’ comments are very different in tone
Commuters
the service
money
– small delays (by minutes, in some cases) have a big knock-on effect
Business travellers
Travelling in comfort
trains (old rolling stock, shabby, dirty, and so on)
catering)
availability online and (b) actually getting the seat they booked
Leisure travellers
helping them, at station and on train
information, announcements etc.
and luggage / storage / buggy space on trains
especially longer journeys
Most negative comments and emotions Elective journeys elicit fewest negative comments and emotions
39
NOTE: Online dashboards allow you to explore the themes in detail, by TOC, journey type, and so on.
The types of comments individual TOCs receive are very different, reflecting their distinct characteristics. Example comparison:
HEATHROW EXPRESS
negative
about speed, efficiency, no hassle VIRGIN TRAINS
(often a negative)
(power/WiFi not working)
lacking, problems with toilets and so on SOUTHERN
cancellations
money
Fewest negative comments Most negative comments
40
People use the ‘any other comments’ opportunity to:
questions (for example, fleshing out WHY they were dissatisfied with a journey)
example, talking about feeling safe not on ‘this’ journey, but on journeys in general (or on other journeys they have had)
invest or do not invest (as they see it), the age of rolling stock, and so on
announcements, hard-to-navigate ticketing, car parking, and more
The comments are a potential gold mine of insight
41
In general
The things that drive trust are the basic truths of travel: on time, reliable, predictability There is no one silver bullet for the passenger experience. It is often an accumulation of small things Old rolling stock, dirty trains and so on feed a narrative of underinvestment For frequent travelers, especially commuters, issues are exacerbated TOCs vary significantly in terms of the issues raised in the comments – there are many individual learnings
At the station
Improve clarity of information and announcements: need to avoid confusion and stress Accessibility is frequently referenced – especially by
travelers with children etc Station facilities, especially shops and toilets, can be improved (also WiFi) Station staff are a great asset, although there are some stories of rudeness etc Commuters / season ticket holders aside, there is confusion over ticket prices – transparency and clarity
On the train
Overcrowding and general misery is a constant refrain of commuters The train environment is noticed by passengers – a smart, clean train with working toilets goes a long way Where WiFi and power is advertised, it needs to work. Key for business travelers People need to feel safe. Guards play a strong role here, as do other passengers Simple things like adequate luggage space, room for buggies, bikes, etc. make a big difference