Presentation on Innovations in London's Transport: Big Data for a Better Customer Service by Andrew Hyman, TFL at HPC and Big Data 2016 in Central London
Innovations in London's Transport: Big Data for a Better Customer Service
1. Innovations in London’s Transport:
Big Data for a Better Customer Experience
Andrew Hyman - Analytics Research Manager
Customer Experience, Transport for London
High Performance Computing & Big Data Conference - February 4th 2016
2. 1. Who are TfL?
2. Act on Fact
3. Case studies
Session Overview
4. Buses Taxi-Private Hire
Coaches Cycles
River Dial-A-Ride
Underground Overground
DLR Trams
Air-Line TfL Rail
Owner and operator of the largest integrated
transport network in Europe
Surface Transport Rail and Underground
5. Our Purpose
“Keep London working and growing to
make life in the Capital better”
Plan ahead to meet the challenges
of a growing population
Unlock economic development and growth
Meet rising expectations of
our customers and users
Every penny of our revenue is reinvested in
running and improving services on the
transport network
6. London is ‘Big’, so our data is ‘Big’, too...
6
Customers are at the heart of our business
Every Journey Matters
13. The “Act on Fact” Journey – revealing
patterns / trends to enable action to be taken
Data
Sense
Making
Information
Intelligence
Story Telling
Knowledge
Take Action
Impact
Inspired by Stephen Few
15. Customers have rising expectations for
personalised journey planning
12 million users visit
tfl.gov.uk every month
16. Our open data feeds whizzy apps that provide
further sources of real time info on demand
17. Data from Oyster and Contactless cards
help TfL understand how people behave
and their transport needs
18. Insight from our data helps ensure we are
prepared for increased demand during events
Hyde Park's Winter Wonderland opened on 20 November 2015 at 17:00.
During opening times (10am-10pm), nearby stations are much busier than normal.
19. We can visualise, understand and look for
ways to influence travel demand
21. On buses people don’t need to tap out
so there is no record of where you get off
22. A customer taps an Oyster card on the
reader, which records the location and
time.
Can we infer the exit point?
Stop
B
Stop
A
Bus events are recorded in the iBus
system and we can match this with our
Oyster data
22
Big Data informs our Bus Network Planning
Working with MIT we built an algorithm we call ODX
that joins iBus and ticketing data together
23. Station Y
Stop X
Stop
B
Stop
A
From the location
of the next tap (if
there is one), we
can infer where a
customer alights
If next trip begins at stop
X, the current segment is
inferred to end at stop A
If next trip begins at
station Y, the current
segment inferred to
end at stop B
ODX enables us to infer the alighting stop
and to build complete journeys
This helps TfL understand how crowded buses are,
plan interchanges and minimise walk times
24. ODX has informed Interchange Analysis as
part of the Better Junctions programme
25. Putney Bridge closed for
emergency repair work in 2014.
Bus services had to stop either
side of bridge. People could only
walk or cycle across.
Big Data / ODX has also helped us look after
customers during major bridge works
27. What Next? Our Future Aims
27
Many more topics and questions to
explore!
Integrating ticketing, bus, traffic congestion,
and incident data for better performance of the
bus and road networks
Developing further personalised services for
those customers who want tailored information
Predicting platform and train congestion at
stations
Understanding walking, cycling and driving
journeys alongside public transport
Using new data mining tools, machine learning
and geo-spatial visualisations to bring data to
life
28. Proof-of-concept real-time crowding
information to help customers to plan travel
57% of customers say they would change their journeys to avoid crowds
Real-time information would be most effective at influencing behaviour
Designs and use of channels are for illustration purposes only
Plan a less crowded journey
before leaving or on route
using Journey Planner
View real-time crowding information and predicted journey
time via information screens at stations