Predicting the Olympics and its impact on Road Safety
Andrew Betteley, Chief Technology Officer — Connected Car Hubio & Daniel Fisher, Head of Data Science — Connected Car Hubio
Insightful predictions can be made by looking at the past. In this blog, we look at producing a working hypothesis on how Olympic events and outcomes may influence both driving behaviour and vehicle crashes. The US Superbowl provides a good comparative event for study, as it attracts over 150 million engaged viewers across the globe and is repeated on an annual basis.
Hubio Telematics Data during the 2016 Super Bowl
Telematics data analysis sourced from the Hubio award-winning Connected Car platform is shown below. The Analysis focusses on braking and acceleration events, comparing percentage difference between Super Bowl Sunday to the immediately preceding and subsequent Sundays in 2016.
As shown, both braking and acceleration drop by 10–20% during the Super Bowl and whilst there is television coverage, the average trip speed increases. These findings are presumably due to far fewer people being on the road, leading to reduced congestion and thus faster average speed and a reduced need to brake and accelerate in an erratic manner.
This observation is reaffirmed when analysing the big data set Hubio has sourced and harvested from the National Highway Safety Traffic Administration (NHTSA).
Our analysis compares 27 Super Bowl Sundays with both the Sunday preceding and the Sunday following, from 1985 to 2014. The results are analogous to the Hubio telematics data presented above, where there is a 10% drop in the accident rate during the game itself.
However, significantly, the first hour immediately following the Super Bowl and television coverage was accompanied by an accident rate increasing by 70%.
Fatal Car Accidents on Super Bowl Sunday
Focussing upon fatal accidents on Super Bowl Sunday in 2014, NHTSA data shows there were 82 fatal accidents, 44 of which were attributed to alcohol which is 54% of the total fatalities. This is in comparison with other Sundays in 2014, where the average number of alcohol-related fatal car accidents is 24.
Similarly, there were 91 fatal accidents recorded during Super Bowl Sunday 2013. Of these, 45 were related to alcohol, which is around 49% of the total fatalities. This compares with an average of 26 alcohol-related fatal car accidents on Sundays in 2013.
Fatal accidents in 2014 and 2013 on Super Bowl Sunday are shown in a geomap below, no significant pattern was found for fatal accidents and is included only for completeness.
This data irrefutably indicates that the combination of alcohol and Super Bowl Sunday plays a role in fatal accidents across the US, although we don’t doubt that other factors also contribute.
Car Accidents on Superbowl Sunday
In California, there were 294 fatal and injury crashes on Super Bowl Sunday from 2009 through 2013, versus 166 for comparable other Sundays. In Los Angeles, there were 77 fatal and injury crashes on Super Bowl Sundays versus 49 on other Sundays. Similarly, there were 40 in San Diego County on Super Bowl Sundays versus 18 on the comparable Sundays.
Further analysis shows that less than half of all crashes on Super Bowl Sundays can be attributed to alcohol. Other factors include driver fatigue, which could be attributed to the late hour of the games finishing, or a combination of distraction, frustration and disappointment due to the scores.
Accidents on Superbowl Sunday for Disappointed Fans
In a radical and significant study, the New England Journal of Medicine noted that the largest increase in non-fatal accident rate occurred in the home states of the losing team, where the number of crashes increased by 68% for all 27 analysed Super Bowls. In stark contrast, accidents rose only 6% in the winner’s state.
Hubio Telematics Driving Data was analysed from Super Bowl Sunday for the losing state from 2013–2016 and shows a 5–7% increase in negative driving behaviour. Negative driving behaviour includes drivers who are hard braking, hard accelerating and excessively speeding — behaviours well documented as being indicative and predictive of a driver who is going to crash.
Quite simply, the disappointed fan is far more likely to drive badly and crash than a fan celebrating a memorable victory!
Conclusion and the Rio Olympics 2016
There is a relationship between the negative outcome of a sporting event on both driving behaviour and the likelihood of a driver car crash. In other words, if your team, player or athlete loses, you are more likely to drive poorly thus having a higher chance of causing a car crash.
Major sporting events such as the Superbowl are directly related to a significantly higher number of car crashes and this is well documented, both here and elsewhere.
As shown in the infographic below, a complex mix of alcohol, driver fatigue, distraction and disappointment are all factors leading to a vehicle crash after a large global sporting event.
Adaptive learning models and predictive analytics within the Hubio platform get to “know” the connected driver and seamlessly predict when there is a problem and when one or more of these factors are being exhibited. The Hubio platform is able to analyse a driving pattern in real-time, detect when something is out of the ordinary for the known driver, and react accordingly.
The Telematics-enabled Connected Car is a predictive paradigm and offers a truly unique value proposition, in that it is able to analyse driving behaviour in real-time within the context of a global sporting events. Based on the analysis, the connected car platform can dynamically trigger an intervention event, reinforcing positive behaviour and eliminating negative behaviour.
The intervention event can be delivered via the infotainments system, which not only leads to an enhanced, seamless and visceral user experience for the driver, but will reduce vehicle-based fatalities and eliminate all the hidden costs associated with a fatal road traffic accident.
Based on the analysis presented, it is reasonable to propose and postulate that “if Jess Ennis does not finish in the top 3 in the heptathlon, this will lead to the disappointed fan phenomena outlined. As a result, there will be an increase in negative driving behaviour and a higher number of car crashes in the UK.”
The impact of Olympics event outcomes and its impact on driving behaviour and car crashes is something we will be actively analysing up at Hubio using our Big Data Connected Car Platform and Dynamic Data Warehouse.
