For many people, the word “trains” still conjures up images of puffing steam locomotives and old-fashioned, wooden station platforms. These people might be surprised to learn that today’s railroad industry is actually one of the world’s most information technology-intensive businesses. Most of this modernization is due to the evolution of big data, a term that broadly refers to the range of new, huge data types that have appeared over the last decade.
For railroads, big data helps to make the entire industry more efficient, more profitable, and safer. Big data analysis influences everything from issues around railroad engineering and operational performance to developments and trends in different lines of business that impact railroad marketing strategies. Read on for a closer look at how big data is helping take today’s railroads into the future.
How do railroads collect big data?
Railroads today collect massive quantities of data using a wide variety of sources and tools. These include AEI readers, GPS tracking systems, many different types of wayside detectors, video inspections, and handheld field tablets used by operators in rail classification yards or other rail facilities. As these data collection methods become more advanced, frequent, and precise, the data gathered grows not only in quantity, but in quality as well.
How do railroads analyze their data?
In most cases, the volume and complexity of railroad industry big data requires specially constructed analytics tools and systems. Accordingly, many rail companies are partnering with technology firms to develop innovative solutions for their data.
One example of a tool specially designed to handle vast railroad databases is the Traffic Flow Analyzer (TFA), which was created for transportation company CSX by global management consulting firm Oliver Wyman. Built as a custom-designed data warehouse capable of archiving three years of train movement records, the TFA uses data extraction and mapping tools to generate such information as geographic pie maps of yard activity, traffic density maps, and detailed reports. Rather than simply modeling rail routes, the TFA stores the actual route taken by each car and each train on every trip, thus allowing the data to be used with tremendous rigor and precision to complete tasks such as engineering studies of bridge loads or ton-mile and car-mile calculations.
Naturally, with this increased focus on gathering and analyzing big data, railroad IT departments are becoming more integral to railroad operations than ever before. For example, the IT department of Union Pacific, which numbers about 2,000 people, devotes roughly half of its staff to managing all the company’s software systems, including the design, development, and building new software products from the ground up.
Two ways the railroad industry uses big data.
1) Predictive analysis and derailment reduction.
A major factor in reducing the number of derailments in the railroad industry is being able to predict when they might occur, and thus prevent them by pinpointing the specific factors that cause them. This feat requires a combination of data-gathering tools and predictive analytics. Tools like thermometers, acoustic sensors, and visual sensors located on rail-car undercarriages paint a picture of track and wheel operation, and indicate any imminent problems, such as wheels beginning to flatten or bearings wearing down. This data is extremely helpful in spotting imminently dangerous situations. However, when combined with predictive analytics software programs, the results allow railroad companies to predict problems days or even weeks in advance. They can then repair or remove defective equipment in a more timely and cost-effective manner.
2) Real-time analytics to minimize disruption.
The railroad industry revolves around scheduling: crews, tracks, terminals, locomotives, and freight cars are all impacted by the precise timing needed to keep everything running efficiently. But given the interconnectedness of railroad operations, things can quickly spiral out of control if there is no real-time system in place to deal with unexpected events, like breakdowns, weather, and accidents. To help keep things running smoothly, railroad companies turn to automatic, real-time rescheduling systems, which bring in raw data from multiple sources to show exactly what’s happening in the moment, and allow surprises to be handled with as little disruption as possible. In their operations, these real-time systems rely on events streaming in from such places as sensors in brakes, bearings, couplings, rails, and switches; GPS units on trains that provide train speed, arrival time, and separation between trains; and radio-frequency identification in terminals and at designated points along the rails.