Algorithm to Data – The Big Data Solutions of Tomorrow

Algorithm to Data is not only a technical term, but also a feature. After all, it's the small features that will make driving even easier in the future. When the car displays in real time all the dangers of the route ahead, such as a slippery road, it is convenient, and safe. Just showing vacant parking spaces in the immediate vicinity is a small but luxurious feature that saves time and saves nerves.


Data Science Professional


Ca. 6 min

Sharing is caring!

Algorithm 2 Data vs Big Data

Regardless of whether it’s a matter of implementing simple customer requests or complex analysis functions: Manufacturers need data from the vehicle fleet. Lots of data. Here lies the great strength of a methodology that generates positive excitement among some, but only a furrowed brow among others: Big Data. Big Data has long since become part of everyday life and, in the public’s perception, hovers between great promises of optimal data collection and the actual, measurable benefits.

Cognizant Mobility’s Analytics Team is also well aware of these difficulties. On the one hand, of course, these lie in specifically evaluating and processing the sometimes enormous volumes of data that Big Data provides. On the other hand, there is the problem of the very specific questions to be answered by Big Data analyses. For this, Algorithm to Data can offer a solution.

Algorithm to Data Visualization
Algorithm to Data: Artistic Visualization of a Big Data Stream

What exactly can Algorithm 2 Data do?

To meet precisely these challenges, there is “Algorithm to Data”, often abbreviated to “A2D”. Instead of collecting masses of data from the entire vehicle fleet centrally for analysis and customer functions, the analyst sends specific search requests or analysis tasks to an individual vehicle and executes them locally. Only the results of these individual queries are returned to the backend. Most vehicle manufacturers refer to this approach as Event Based Data Collection, or Campaign Management.

This methodology offers numerous advantages over a centralized Big Data solution:

  • The amount of data to be sent is small
  • No high-end network coverage is needed
  • The computing power required for handling the data decreases
  • Individual events can be addressed in a specific and coordinated manner
  • Only data that is actually needed is collected

In A Nutshell

  • Big Data Analytics
  • Algorithm to Data
  • Telematics
  • AWS (Amazon Web Services)
  • Microsoft Azure
  • Spark
  • Scala
  • Python

In addition to these very practical advantages, Algorithm to Data also offers legal security. The question of who actually owns the collected data can therefore be kept simple: The driver. Its data remains on its own vehicle when analyzed via Algorithm to Data. Only the analysis results and the resulting findings go back to the manufacturer, and this data can be easily anonymized.

Algorithm to Data Driver Data
With the algorithm to data method, the data stays where it belongs: With the driver.

How Algorithm 2 Data strengthens thought processes

Of course, it must still be clear in advance of the analysis what data is to be searched for in the first place. Thus, Algorithm to Data definitely forces users to think about what the goal is even before the analysis begins. One advantage is that it is possible to dispel the widespread, albeit often unfounded, premise that big data merely collects massive amounts of data, and that users no longer know which data is relevant.

However, even the most practical approach often has a hidden disadvantage: The current hardware is not always suitable for targeted analyses on a large scale, especially if these are to function via individual controls. A significant leap in the performance of current computing components still has to happen here. Also, data can hardly be analyzed retrospectively. Manufacturers are therefore relying on so-called connected data recorders in the development phase. Powerful computing hardware, which is additionally installed and not only guarantees the traceability of the data, but also has the purpose of detecting errors or conspicuous behavior of the software already in the vehicle and thus to pre-filter it. Over-the-air updates can be applied directly to the vehicle after errors have been detected and eliminated.

Cognizant Mobility and Algorithm 2 Data

Cognizant Mobility works with manufacturers to design and implement data and analytics strategies. Not only for pre-development, but also for series production, resulting in customer functions. This includes precisely those software modules that are relevant for data exchange with the vehicle.

The Big Data team at Cognizant Mobility also takes care of selecting the cloud architecture required for this, as well as the subsequent analysis with automatic pattern recognition. In most cases, Amazon Web Services is used, as well as the Apache Spark framework and the Scala programming language.

After all, it’s the small features that will make tomorrow’s driving even easier. And the Cognizant Mobility provides not only Algorithm to Data, but also many other solutions.

Professional Daniel


Data Science Professional

Dr. Daniel Isemann ist Head of Data Science and Artificial Intelligence, dessen kluge Lösungen er seit Jahren in den Dienst der Cognizant Mobility stellt.