Big Data, a data future built on a blank slate
Big Data is everywhere, in blogs and in newspapers. It is the topic of the moment.
Looking ahead to #meetPSAexperts on 3 December 2014, “Big Data: imagining a data future”, I talked to Jean-Pierre Dumoulin, expert in the operation and infrastructure of Information Systems. I asked him a few questions about Big Data and how it is set to change our lives.
How would you define Big Data?
Jean-Pierre Dumoulin: To my mind, Big Data is made up of two things. First, the technologies we use to capture, collect and store very large quantities of data at an affordable cost. And then the tooling that we use to structure, analyse and apply mathematical models to these data in order to create value for the company and its customers.
To illustrate this concept, let’s take Google Translate, the well-known translation tool from the Internet giant. Contrary to what you might think, it is not based on large dictionaries and artificial intelligence. It is based totally on Big Data! Google has harvested translations available online around the world and applied statistical models to them in order to find the best translation for each word, depending on the context, from one language to another. This makes it possible to translate texts into many different languages. Above all, quality is improved continuously as the system acquires more translations and corrections. Big Data therefore involves the capacity to harvest large quantities of data for an intelligent purpose. This requires not only IT and mathematical tooling but also human skills, provided by the data scientist.
What are the applications for a vehicle manufacturer?
JPD: I would say that companies can use Big Data in four ways:
• Create value with data already existing within the company: using bases of structured data (around 10%) and also non-structured data (90%) to make enlightened decisions, based on more reliable and exhaustive data. For example, Big Data can contribute to cybersecurity. By using the data collected on IT systems, you can identify anomalies and thus any flaws. If the same employee is logged on in France and in India at the same time, then you have a problem!
• Improve customer relations: with the current survey system, when you talk to a customer, you only get answers to the questions you ask. With Big Data, you can harvest all the public information on line (from the social networks, for example) grouping everything people have said about a brand. This vast mass of online data can be used to identify trends that will enable you to answer customer requirements more effectively.
• Improve products: sensors fitted in the engines make it possible to conduct preventive maintenance and so to anticipate failure, but also to better understand the real behaviour of an engine in order to tailor the design process to real conditions of use. At PSA Group, for example, the data provided by the sensors on test vehicles provide a clearer picture of the way vehicles are used by real customers (and not just the customers imagined in the mission brief). The vehicle sensors can be used to obtain anonymised data in real time from customers who agree to give us access to their data.
• Develop services linked to connected objects: in the case of cars, vehicles pass on and exchange data via connected functions. This makes it possible to create new services. While maintaining a totally ethical approach, you could, for example, offer services such as personalised maintenance plans for customers, predict the failure rate of parts, give driving tips or automatically book an appointment with the dealer. Connected vehicles will be a major differentiating factor because, with 600 million vehicles on the world’s roads, they provide a wide viewing angle on an object with real social value and high visibility.
Why is PSA Group interested in this subject?
JPD: At PSA Group, a recent vehicle has 60 sensors supplying 4,000 items of recoverable data. The IT network links all this in-car data through a wireless connection. This network is now connected with the outside world, making it possible to invent new services for customers. Our main concern is ethics. We never collect data without the customer’s knowledge. If an item of data is anonymous (such as the engine temperature), it is accessible, but if it can be traced back to the customer, then it is private.
Imagining a data future: in 2020, Big Data will be behind a number of revolutions such as…
JPD: I believe that the revolutions driven by Big Data will be:
• A completely transformed business model for many companies. Data is the black gold of the future! Companies able to take advantage of all the Big Data technologies will have a far higher level of expertise with respect to internal performance and customer knowledge. Amazon, for example, has far more information on its customers than any other business in the world, and an incredible qualified database of customers on international markets. If Amazon suddenly decided tomorrow to sell insurance, it would be able to do so without any intermediaries. Amazon could easily calculate risk by cross-referencing all the information in its possession and reach the customers that already trust it. Big Data will be a real differentiating factor in building competitive edge.
• The invention of new and incredible services: with billions of sensors, we could create new services! Existing GPS systems are based on updated maps. Tomorrow, we could use all the sensors from connected objects to update the maps in real time.
• We’re starting off from a blank sheet, but the techniques are improving every day. IBM’s “Watson”, for example, can make automatic medical diagnoses. It’s up to us to invent the right uses and to aim for a future that is better rather than worse than before!