Who wouldn’t love to be able to predict traffic patterns? That’s exactly what BMW Group is doing in Munich in conjunction with Splunk’s Machine Learning Toolkit.
During Wednesday’s keynote at Splunk .conf18 in Orlando, Fla., Boulos El-Asmar, a research scientist focused on machine learning with BMW Group, shared how he and his team are using data analytics to optimize the company’s test fleets.
“Today, we are getting so much value from our data,” said El-Asmar of the effort, which is part of a project through Splunk’s Machine Learning Advisory Program. “We’ve been working around predicting traffic in the city of Munich. To be able to predict when and where car-driving events start in the city can be of high value to carmakers to be able to monitor their test fleets or public transport systems.”
From the get-go, he says, BMW Group focused on designing a solution that makes it easy to incorporate multiple sources of information, such as weather data or public transport data.
“Or events data, perhaps Oktoberfest, which I’m actually missing to be here with you today,” El-Asmar joked.
Visual Views of Actual vs. Predicted Dynamics
Along with Splunk’s native features, BMW Group was able to build its solution and create reliable models that predict traffic in real time, El-Asmar said. Given one month of training data from its test cars fleet, it was able to predict for the following week car-driving events down to every hour of every day.
“Splunk gives us visual views of what actual versus predicted dynamics would look like,” he said. “Machine learning can sometimes be intimidating, but MLTK is really easy and straightforward. The guided workflow has made it easy to get data and import and train our custom machine-learning models.”
Prior to El-Asmar sharing BMW Group’s successes, Splunk announced the launch of its Machine Learning Toolkit 4.0, which it touts as having more scale and improved collaboration with an open source community.
The company also unveiled a variety of offerings under Splunk Next, including beta availability of Splunk Business Flow, which provides cross-session and cross-channel visibility for users, and Splunk for Industrial IoT, which features real-time monitoring of industrial equipment, detection and mitigation of operational technology security risks, and condition- and predictive-based maintenance.
“Using sensor data and machine learning, we’re going to help the factories in industry 4.0 get even smarter,” said Seema Haji, director of product marketing for IoT at Splunk.
Additionally, it announced the launch of Splunk Natural Language, which allows users to query Splunk using natural language or voice text.
Keep this page bookmarked for articles from the event. Follow us on Twitter @BizTechMagazine, or the official Splunk Twitter account, @splunk, and join the conversation using the hashtag #splunkconf18.