The International Light Rail Magazine
+44 1733 367610

Big data, big opportunities

Predictive analysis methodologies utilising complex and diverse data streams allow operators and authorities to more accurately forecast demand, manage disruption and allocate resources within multi-modal networks. This is Algiers, which features metro, tram and bus operation and hundreds of connection points. M. R. Russell

The transportation industry has quickly become one of the leading players in the connected world’s Internet of Things (IoT), because of the extensive data generated each day by vehicle locator and passenger counting systems, as well as ticketing and fare collection systems.

While already consuming vast amounts of storage space in operators’ data centres, driving value from these mountains of information has proved elusive. Traditional analytical techniques cannot be used due to the scale, complexity and disparate nature of the data.

Fortunately, advances in data science and analytics are creating an opportunity to harness this information for new insights into passenger journeys to optimise transportation networks and rider convenience.

Deriving value from public transportation data starts with a prominent issue: all of this data is housed in separate silos, generated by different systems from different vendors in different formats.

By looking at this information as singular categories, however, transportation agencies and authorities are viewing their networks’ realities through a small pinhole.

The opportunity is to see a more complete, richer picture of the factors and dynamics that contribute to the success, effectiveness and efficiency of the network by fusing these data sources together.

The opportunity is to see a more complete, richer picture of the factors and dynamics that contribute to the success, effectiveness and efficiency of the network by fusing these data sources together.

Analytics companies, like Urban Insights, can help realise that opportunity by leveraging big data tools and predictive analytics to assist agencies in improving operations, reducing costs and better serving their riders.

How trips become ‘journeys’

To better visualise the types of data transportation authorities and agencies are fusing, let’s envision an everyday scenario: you board a bus at point A to travel to point B, and at point A you touch your smartcard – creating a data footprint.
In most transit systems, you are not required to touch your smartcard to exit. Perhaps the next trip entry is at point C, with a travel segment from point C to D.

But point B, and whether it indicates a transfer or mode change from bus to light rail, for example, is getting lost, and there’s nothing to record location D. Planners are then left guessing about the journeys riders are taking.

To compose trips into journeys, big data and predictive analytics can combine and analyse multiple types of data in order to understand the progression of travel and develop a deeper understanding of traveller needs. Why is it that someone chose a particular route, based on the journey they were actually taking?

The predictive analytics process

Success in achieving a fused view of activities taking place in the transport network is reliant on leveraging knowledge of existing systems, incorporating specific factors relevant to a particular transit organisation. Collaboration in this way allows organisations to gain impactful analytic insights in a quick and organised way.

The first step in gathering the relevant data is extracting and collecting it, to survey the available sources and extract it. Then it must be stored in a staging area prior to being cleaned and organised into standardised transportation-specific models suitable for analytics.

Another critical step in the process is to interpret the story told by the collected information. This can be achieved by fusing domain-specific data with other modelled data and applying predictive analytics to create a more complete picture.

To exploit these relationships, statistical analysis, simulation and optimisation can be applied. This will help to plan for and predict what’s likely to happen under different scenarios; these gathered insights can then be turned into data visualisations and reports to convey nuanced behaviour and trends.

The last step in harnessing insights for operational improvement is for transport organisations to realise the true benefits by achieving their desired and planned operational outcomes.

The future of transportation analytics

The number of transportation authorities taking advantage of these opportunities will continue to grow – and the technology will evolve along with them.

In the future, specific tools and models will work with data sources from an increasing number of operational systems as well as non-transit sources such as traffic data, social media feeds to reveal traveller sentiment, population demographics, geospatial data, weather data and economic and retail data to improve operational management and planning activities for transportation.

Agencies, for example, may use big data to predict the impact on highways and public transport networks caused by metro line closures or roadworks, and recommend the most effective change in communication strategy and scheduling to mitigate some of the impact. The same concept can be applied to unplanned events, such as transport labour strikes, or on a more regular basis to manage traffic incidents and congestion.

The future of data analytics in transportation offers many opportunities. The challenge is certainly not the ability to generate data. The solution is pushing forward using significantly improved means and methods to gather and understand the available data so business decisions can be informed by better comprehension of what the data is saying.

For the full version of this feature including case study, please see Tramways & Urban Transit – October 2014 issue (page 431).