The Amazing Ways Big Data and Predictive Analytics Can Reduce Traffic Casualties

by   |   January 23, 2017 5:30 am   |   0 Comments

Bernard Marr

Bernard Marr

Lots of people are scared of flying, but it’s rare that you find someone who is scared of getting into a car. This is strange because statistically it’s one of the most dangerous things you can do – in the US alone, 35,000 people were killed in traffic accidents last year.

While this figure has been steadily declining from its peak in the 1960s thanks to improvements to automobile safety, it still represents a huge waste of life. Accidents also put a huge strain on emergency healthcare resources and cost a great deal of money to administer.

However, several initiatives involving Big Data and analytics are now being put to work to solve the problem. Policing, speed-reducing road furniture and driver education programs have all been shown to play their part in reducing casualties. The problem is knowing where and when to deploy them, and today’s Big Data technologies mean we are getting better at predicting this.

Working with IBM, the Tennessee Highway Patrol (THP) has implemented a predictive modeling solution which uses data such as previous accidents, DUI (Driving Under the Influence) arrest statistics, weather, and events data to predict where crashes are likely to occur. This allows them to deploy resources such as patrols and speed-calming measures and has led to a 6% cut in casualty rates and a 43% increase in DUI arrests.

New York, where someone is killed or seriously injured by a vehicle every two hours, is one of the cities signed up to the Vision Zero initiative, which aims to cut the number of traffic deaths in targeted cities to zero. As part of this, a collaborative project involving Microsoft and data scientists from DataKind includes using Internet of Things technology such as cameras and sensors to enable more accurate modelling of traffic conditions. Alongside this, the initiative leverages data from the Department of Transportation and the city’s open data. The aim is to develop a more accurate, data-driven understanding of what interventions have the biggest impact on the number of people killed or seriously injured by traffic.

These interventions can be in the form of engineering (building safer roads), education (teaching drivers to use the roads more safely) or enforcement (punishing those who contribute towards making roads dangerous).

In particular, a variable known as “exposure” – the total number of vehicles, cyclists, and pedestrians in the vicinity of an accident – was particularly difficult to understand before the emergence of predictive modelling and IOT technology. Now there is hope that today’s advanced analytics tools will bring about the insights needed for the job – early results are said to be encouraging and the pilot has been extended into three other US cities.

Outside of big cities, and particularly in remote areas, hazardous weather conditions are a big contributor to accidents and road casualties. The Kentucky Transportation Cabinet is responsible for keeping 28,000 miles of highway clear of snow and ice which can pose a danger to drivers. There are strict targets that it must meet when it comes to coordinating its fleet of 120 snow removal trucks to keep the roads clear of snow, and pressure to spend the public money involved as efficiently as possible.

The solution they chose was to design a Hadoop-based analytics system designed to read data from sources including the HERE network of traffic sensors deployed throughout the state, social media posts, and Google’s Waze GPS and traffic information service. The system is able to optimize use of materials such as salt used to remove ice and prevent accidents, minimizing cost as well as environmental impact.

Of course, autonomous and self-driving vehicles have long been heralded as the tech quantum leap which will finally make travelling by car as safe as any other method of transport. These rely on a huge amount of Big Data-driven analytics to operate safely. Each vehicle is essentially an entire array of moving sensors, designed to capture and process every bit of data that they can from the vehicle’s environment. They also pose new challenges, however – particularly concerning how they will cope and interact with unpredictable and erratic human drivers once they are on the roads in large numbers. However as human error is said to be the primary cause of 94% of automobile accidents, removing that from the equation should, most experts agree, go a large way towards improving overall safety.

Traffic accidents are the products of bad roads and bad drivers. Undoubtedly the data which is needed to assess how the balance of these factors contribute to real-world accidents is out there, but until now it just hasn’t been affordable or practical to capture it. Thanks to the rapid advancements in sensor technology, and analytics to make sense of the sensor data once we have it, the roads will hopefully soon be far less dangerous places.

 

Bernard Marr is a bestselling author, keynote speaker, strategic performance consultant, and analytics, KPI, and big data guru. In addition, he is a member of the Data Informed Board of Advisers. He helps companies to better manage, measure, report, and analyze performance. His leading-edge work with major companies, organizations, and governments across the globe makes him an acclaimed and award-winning keynote speaker, researcher, consultant, and teacher.

 

Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise, plus get instant access to more than 20 eBooks.

Tags: , , , , ,

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>