Big data as life savior! Read how these data analytics ideas are preventing accidents

Accidental risk prevention using big data

Big data is transforming the way accidental risks are predicted

With predictive analytics at the core of big data, it was only a matter of time before we would have seen big data predictive models being deployed to prevent or assess the risks of accidents. What is also important to note is that this field, but from interventions from IoT tools, isn’t entirely new. What big data has achieved, in comparison to conventional pen-paper and computational tools, is reduction in long-term costs of data collection, ability to process massive data sets at unprecedented levels to increase accuracy of prediction and most importantly, real-time analysis. In the recent past, several innovative business ideas have hit the market. Some of them have been innovated in academic quarters, some others are product of the startup ecosystem and interestingly, some of them have also been developed by city administrative units, who are always striving to make city roads safer for commuters. Insurers, industrial plants and city administrative units form the major consumer segments for such products. In today’s article, I have looked at half a dozen of such noble initiatives that are building life-saving mechanisms for us.


Utah WVC Reporter

Utah WVC Reporter was incepted as an academic project. Led by Dr Olsen at Utah University, his team developed a smartphone app called WVC Reporter to prevent Wildlife Vehicle Collisions (WVC) in Utah. Was this app needed? The definite answer is yes. According to a published report on the US market, the annual damages from WVCs are in excess of $8bn and results in more than 200 deaths each year.

The app sources wildlife species geo-location and details via a crowdsourced smartphone content route, which also allows for offline data entry from phone users. While the user manually enters the wildlife species’ identification data when she spots it somewhere, geo-location data is automatically gathered by the app. This data is then uploaded in real-time to the central database, also accessible via web, and can be then accessed by interested stakeholders in GIS compatible file formats. Needless to add, this method has distinct advantages over the conventional pen and paper methods and deserves scaling up.


Synaptor

Synaptor is a Western Australia based startup that runs data analytics on incident logs and other site reports generated at mining and construction sites, creating real-time interactive risk maps. These maps can be used to predict the probability of an injury or a hazard and generate timely alerts. Synaptor has already signed up big ticket clients such as Colgan Industries and Airpac Bukom.


Virginia Tech Smoke Words Study

In 2012, Virginia Tech researches conducted a study to analyze if mining of social media and automotive forums for “smoke words”, words that could suggest automotive defects in posts, could result in early detection of assembly-level defects in automobiles. Apart from potentially reducing the expensive cost of recall for manufacturers, the success of this approach can prevent any accidents that are caused because of undetected manufacturing issues. In simpler terms, this is the automobile equivalent of the Google Flu Trends. This is not as simple as it sounds though. Detection of “smoke words” is not easy as they are not characterised by conventional sentiment data analysis methods. But the fact that the utility of such big data analytics tool can be expanded to industrial or occupational health and safety setups makes a strong case for concerted effort to its commercial product development.


Boston’s Street Bump App

Boston city has developed a mobile app called “Street Bump” that identifies potholes on city roads. The app uses a smartphone based sensor technology that detects slight changes in the phone’s accelerometer when a commuter driving her car on the road, hits a pothole. The data is then collected from user smartphones and a real-time database is created for the municipality to act upon for road repairs. Compared to the currently employed manual surveys that involve engineers running truck-chain systems on the roads, this approach is a steal – particularly from cost savings perspective. Not only will this method save costs of annual surveying, but also timely repair of the potholes could shave off some pie of the annual $6.4 billion that potholes costs to the US drivers in road accidents.


Bridgecrest Medical

I had showcased San-Francisco based startup Bridgecrest Medical recently in another article titled Unique healthcare big data startups you should definitely know about for its successful Ebola prevention efforts in Africa. Today, however, we talk about its flagship product “Fatigue Management Solution” meant for prevention of industrial and mining sites employee accidents. This tool collects employee health data from conducting their health screening tests for ailments such as sleep apnoea and making them wear wearables to constantly measure their vital health statistics, sleep patterns, etc. Both these data sets are maintained in a big data repository and analysed against control groups and recorded accidental risks datasets to identify fatigue levels of employees. The tool then classifies employees in high, medium and low risk categories. The expertise in this analysis is brought to the table by empanelled medics who work in leading global mining companies.


Honda Virtual Tow

In a motor event at Detroit in September 2014, Honda showcased its Machine-2-Machine technology product series. A blend of IoT and automotive engineering, it displayed two connected Acura Sedan cars “talk” to each other – enabling a “virtual tow”, which allows a driverless car, directed by the other. According to Honda, the virtual tow technology has been designed to help drivers assist each other in case of breakdown. In addition, Honda also displayed technology that integrated rear-view camera data with IoT platform, enabling cars to analyze hazards on roads and switch lanes, if necessary. At the event, it also revealed vehicle-to-bicycle and vehicle-to-motorcycle technologies that would timely alert a motorbike or a smartphone about an imminent collision. Maintaining the big data over a cloud, Honda plans to play a pivotal role in future automobile space dominated by big data and IoT.

Anubhav, a data scientist, writes about new developments and future trends in the machine learning and data analytics domain.
He can be reached at anubhav@thinkbigdata.in
Follow him on Twitter at: https://twitter.com/think_bigdata

Leave a Reply