Combining previously unrelated sets of data can provide an in-depth view of travel patterns.
"Through the use of analytical tools,
While an authority may have a large body of data it is usually collected and stored separately in ‘silos’ with little or no correlation or coordination between the various data sets. Many times the data gathering is set up to ful l one stakeholder’s requirements; other stakeholders have different requirements and therefore create parallel but often unconnected detection and monitoring systems. According to Rosado, many organisations do not have the full knowledge of all the data they are collecting.
“Many times these discreet systems are designed by different companies and are not built to communicate or naturally link with each other; we aim to break down these barriers and release the data from the silos,” he says. By combining and analysing these data sets he says Urban Insights can derive an array of meaningful information to support an authority’s decision-making.
He cites San Diego Metropolitan Transportation System as an example. The authority runs light rail and bus services and while it has a single ticketing system it does not have visibility into transfers between routes and modes, as each recorded transaction is for an individual ride. Most acute is the trolley service because it is an open system without barriers and works on a trust system with passengers able to validate electronic tickets on the platform. Paper tickets are also available from vending machines and while the vehicles are fitted with GPS location equipment and automated passenger counting (APC) equipment there is no direct link between the systems.
“The APC records are time stamped and show how many people boarded and alighted from the trolley at any one time. But as it doesn’t cross-reference the timetables, there is no record of speci cally where the passengers board or alight, let alone how many had valid tickets for the journey,” says Rosado.
By using algorithms to associate the GPS-corrected timetable routing with the time stamp on the APC data, it was possible to get a much clearer view of how many people were getting on and off at each stop. This was done across the whole network and averaged by time to show the typical number of people boarding and alighting at each stop by both time of the day and day of the week.
This detailed information can be used for planning purposes to answer some speci fic questions - but it is not the complete picture. “It’s no good looking at the network as a series of individual trips because that’s not the way people travel and does not re flect why they are travelling. They didn’t choose to go to ‘Old Town Station because it fulfils some need they have, it’s just on their way to somewhere else,” Rosado says.
Such detailed information made it possible to determine if travellers were using the transit network in the way the planners had envisaged and if the planning assumptions behind the routing and timetabling are consistent with travellers’ needs. It also provided a more complete understanding of the proportion of travellers who may not have a valid ticket and where and when the largest proportion of non-compliance occurs – allowing enforcement or education campaigns to be better targeted. However, the exercise undertaken by the MTS was aimed at determining how it can better meet the needs of travellers and to ensure its services are optimised to those needs.
“With the legwork done to create the association between travel activities and the service offered, it is possible to look at the usage in comparison to the network configuration to see if there are inconsistencies,” says Rosado, adding: “That’s the phase we are in right now evaluating if there are locations and periods of the day when the usage is incongruent with the service offered.”
Similarly with bus services, while passenger ticketing transactions were time stamped, they were not geo-located and so did not provide a full picture of which sections of the route were busy or quiet. And while driver feedback is available that may not be a sufficiently robust basis for restructuring a service. “Planners would probably look to have more concrete data to inform those decisions and often this comes through a survey - but that is a sampling process producing a single snapshot in time,” says Rosado.
In San Diego’s case the primary aim for the new analytics was to identify where travel patterns were inconsistent with planning assumptions – for instance if more passengers than expected are transferring services at a particular point or at a particular time. If this occurs the authority could consider adding a new service, or increase the frequency and synchronise services at particular times of the day. However, it is currently too early for the San Diego information to have fed through into revised routes and timetables.
With other authorities both the requirements and data sets will be different and over time Rosado believes the use of such analytic techniques will expand to answer many different questions.
While Urban Insights’ staff can devise the analytics and highlight anomalies, it is the authority’s staff that will interpret those results and decide if action is required and what changes are necessary. As Rosado puts it: “We implement a process that provides authorities with the information they need to support the decisions they are required to make. We also make recommendations in light of the authority’s objectives based on what we see in the data and at the same time collect their feedback to constantly improve the services we offer.”
As this expertise is incorporated into the system it will be possible for an authority to model new routes, additional services and timing changes using its own data to evaluate the effects of planned changes before decisions are finalised. “By examining options such as consolidating services at particular times on certain routes, it may be possible to make changes which would both reduce operating costs while also improving the travellers’ experience,” says Rosado.
He is keen to emphasise that Urban Insights’ analytical techniques are equally applicable to other forms or transport, including multimodal, and to authorities worldwide. “If you really want to balance the demand with the capacity of the transportation network, it has to be optimised across all travel modes.”
There are also occasions where the analytics can inform urban planning and land use decisions. “Existing transport modes have a finite capacity and identifying unused capacity can show where additional growth can be accommodated without needing to expand the transport infrastructure.”
When asked how long it would take to carry out such analysis Rosado says: “We aim to bring this down to weeks or a few months. One of the challenges that has hampered the use of data to support decision-making is that often by the time the information has been collated and analysed it is no longer relevant or the authority has already had to make the decision.
“While the likes of TfL or NY MTA have teams able to process this type of data many other organisations will not have the IT systems or the expertise required for this type of analytics. Urban Insights can offer them that capability without the need to build teams or install new IT systems.
“These capabilities are designed for repeated use in a constant improvement cycle, where it’s necessary to measure performance against targets and benchmarks, to determine if an initiative has achieved what was intended.”
Regarding a typical payback period, he says that is related to the time required to implement changes in the transport network. Now that is another question altogether.
Corporate structure
Urban Insights a subsidiary of
