We need data to make decisions about the effectiveness and efficiency of safety and other infrastructure improvement programs based on data. But we are consistently confronted by questions about data.
That could be the opener to many a mid-market company internal memos. Many of those companies are working on getting the right culture and infrastructure to transition to data-driven decision making. Well, except for the safety part. Typically in its place you would see something like sales maybe.
Some of the questions the RSDP grapples with have to do with the type and quantity of data to collect, the parties responsible for data collection, data management and integration, sharing data among departments/agencies, and the type of analysis to perform and tools to use.
These are just some of the big questions, and the FHWA is not the only one dealing with them.
The FHWA has some technical assistance programs designed to help agencies improve the quality of their roadway data. But is seems like the challenges are not met entirely in-house for Departments of Transportation (DOT) across the US. This is where 3rd party actors like Numetric come in.
Numetric's mission is to empower traffic safety agencies to be more efficient and data-driven to make roadways safer. Numetric offers a suite of applications, as well as data preparation and data management services. Numetric works with many DOTs, and today is releasing an update to its Safety Analysis application. The application, Numetric says, reduces cost and time state and local agencies incur to make roadways safer.
The HSM contains concepts, guidelines and algorithms for analyzing crash frequency prediction and is incorporated into every state's roadway planning, design, operations and maintenance decisions. In a Q&A ZDNet had with Danny Anderson, Numetric VP of Product, we asked about the HSM and how it informs what Numetric does.
Anderson said that the engineers who developed HSM in the traffic safety space are always looking for improved ways of making the roadways safer, and Numetric uses these guiding principles of safety in its analysis. He pointed out that Numetric's analysis does not only create crash-based analysis, but systemic analysis as well:
"Much of the HSM is based in crash-based logic, understanding why specific crashes occurred. What systemic analysis provides is the insight into what roadway types are potentially problematic. Instead of waiting for crashes to occur, systemic analysis can predict what will occur. By having both forms of analysis, Numetric users have a greater depth of understanding around not only what has occurred, but why it has occurred".
Moving up the analytics stack
Anderson said that Numetric is not tied to HSM guidelines, as organizations have the ability to replace calculations or logic based on what works for them. Numetric, he went on to add, brings all types of data together (spatial, relational, full-text, time-based) in its cloud-based platform. Data integration is one of the key challenges, and it enables enhanced searchability, enforcement of greater data quality and consistency.
This is all in line with what we know: investment in data literacy and infrastructure is a must in order to reap the benefits of data-driven decision making. It sounds like by effectively outsourcing their data processing needs to Numetric, DOTs are moving up the analytics evolutionary chain: from descriptive to diagnostic and predictive analytics.
But there are bigger questions to ponder here: is outsourcing really the way to go for road safety data? Is this a future-proof strategy? Should not agencies have their own infrastructure, and develop their own data-driven culture?
Anderson believes that initiatives like the FHWA's Data-Driven Decision Making work to help states start using data, but "for groups that struggle with using data to make decisions, once data coming from the digitization of vehicles become more widely available, these groups may be so far behind they won't be able to catch up".
Numetric argues that agencies are taxed by spending scarce time and resources to diagnose the root causes of crashes and try to create effective solutions to help remedy those crashes, a process that is a critical part of the safety management cycle. As a result, safety engineers and designers often are left with having to prioritize roadway projects for evaluation since this arduous process consumes precious resources.
"Unfortunately, some agencies have no other choice than to proceed with new projects without a safety report because of limited resources. However, with this new app, agencies no longer have to choose which projects will or won't receive this critical analysis. This app is the first of its kind and the only one available to actually present suggested solutions", said Anderson.
He went on to add that the data, or access to this data, is provided to Numetric by agencies who choose to partner with them. Agencies retain the rights to their data, and Numetric works with the agencies to meet all criteria for handling data, including PII data.
More data, yes, but how?
Truth is, we don't know enough about the internals and the economics of DOTs to be able to have an informed opinion as to whether outsourcing is the right choice. Whether to outsource or not, what and to what degree, is always a challenging decision. The fact that this is public data, taxpayer money, and a sensitive topic, don't make it any easier. In any case, it's worth examining other approaches on the topic as well.
Some new approaches for data-driven and smart road safety policies in cities and regions were presented in 2018 by the likes of the NYC DOT, Transport for Greater Manchester, the European Transport Safety Council, Volvo and the SharedStreets & Open Transport Partnership. There was an emphasis on limiting vehicle speed in cities, as there is a link between speed and fatalities. But more fundamental questions were raised, too.
POLIS, a network of European cities and regions cooperating for innovative transport solutions, drew the link between safety and sustainability. The argument is clear: to improve road safety, reducing vehicles is an obvious avenue to pursue. Safer roads, cycle lanes and footpaths will encourage people to walk and cycle more. Fewer motor vehicles make walking and cycling safer for vulnerable road users such as pedestrians, cyclists, and motorcyclists, which make up for 68% of urban road fatalities.
This approach questions the use of more technology, and more technologically enriched vehicles, as the single answer for better road safety. The POLIS paper on vehicle automation [PDF] for example questions often unchallenged arguments and assumptions (optimism bias) in the vehicle automation narrative, and embarks on a sober exploration of the real impact of automation on mobility on cities and regions.
Public-private data partnerships for road safety
Anderson, on his part, said that the widespread use and focus on autonomous vehicles will change the way agencies, communities and citizens think about safety. His view is that the need to become more data-driven in safety analysis will become even more critical, and to prepare for this, groups need to understand and use the data they currently have available.
SharedStreets is working on this, too, in its own way. SharedStreets is a project of the Open Transport Partnership, a non-profit organization that builds tools for public-private collaboration around transport data blending technology and policy. SharedStreets is building software, digital infrastructure, and governance models to support new ways of managing and sharing data.
SharedStreets is running a number of pilot projects powered by collaboration between cities and private sector innovators. These projects aim to produce standardized and replicable data models and reusable open source tools that help cities and of all sizes follow in others' footsteps, and build crucial partnerships between the public and private sectors.
This seems like a more sustainable model in the long run: empowering organizations to be as self-sufficient as possible, rather than outsourcing critical assets and know-how. Ultimately, this raises questions of governance and control in cities and regions that go way beyond data. Perhaps the most (in)famous example is Uber, and its numerous disputes with city governments.