To immediately avoid setting off down the wrong track, the term “Big Data” has, since its emergence, outgrown its simple meaning. It will always be considered the ensemble of the technologies and analysis methods for large amounts of data.
One universally-accepted definition of Big Data does not exist; this is simply due to the fact that it has always been an evolving methodology, whose boundaries expand just as quickly as the quantity of data to be analysed does. To understand what we’re referring to, let’s consider that data traffic from the Internet alone developed 800 Terabytes in 2000, 160 Exabytes in 2006, 500 Exabytes in 2009, 2.7 Zettabytes in 2012, and it is predicted it will reach 35 by 2020, according to an estimate that today one might even say falls short.
Facebook alone generates 10 Terabytes of data each day, and Twitter 7. But social media is only the tip of the iceberg.
In the first instance, we can accept the definition that “'Big Data' refers to datasets whose size is beyond the ability of typical database software tools to capture, store, manage, and analyse” (McKinsey Global Institute, 2012).
The Big Data phenomenon began to appear substantially around 2012, at the same time as the main changes to the world of Information Technology and the birth of large Data Centres that led to:
• the increase in availability and capacity of storing;
• the rise in computing power;
in parallel with a growing availability of data thanks to the generalised connectivity between things and people.
These three factors steered research towards systems and programmes to collect and analyse a large amount of data, researching tools and innovative methods, transforming Big Data from a phenomenon to observe and describe, into an applied methodology.
The industry sectors involved include... everyone!
Every sector, from manufacturing and the automobile industry to agriculture and services, has data, images, video, audio, numbers, trends and behaviours to be analysed, cross-checked, transmitted and made profitable.
From the start, the OTT (over the top) strategy of big players like Google, Amazon and Facebook, was to buy as much data as possible, primarily through their own platforms, then with the acquisition of companies and start-ups that had already begun to buy Big Data, to manage and control them. It was also done through Apps (e.g. people’s movement with Waze GPS) and tools that can collect valuable information about customers (e.g. Nest thermostats).
Nowadays, large million-dollar acquisitions – in 2013, Google bought Waze for 1.1 billion dollars, on top of 3.2 billion dollars for Nest and 1.65 billion dollars for YouTube – that apparently cannot be explained from an immediate profit point of view, are no longer simple bubble speculations for investments, but appear to be decisively forward-thinking strategic decisions.
Reaching, winning over and keeping the customers whose 360° behaviour history you own, is undoubtedly the killer application that cuts across all industry sectors.
From the first model of the 3 Vs, to life cycle... the process and analysis take shape.
The first Big Data analysis model could be considered “The 3 Vs” named by Douglas Laney, where the “3 Vs” used to characterise Big Data (1. Volume – 2. Variety – 3. Velocity) are now five, thanks to the addition of Veracity and Value.
Almost immediately, in fact, it was realised that compared to Big Data, Volumes (large quantities), Variety (different kinds of data) and Velocity (of generation and delivery), are useless parameters if we cannot also measure Veracity and Value.
The availability of data is now certain and for some time companies have been noticing its value, but only recently have they focused on the fact that these data can truly be transformed into a final or intermediate product in a monetisable value chain that in part, still needs to be defined.
It cannot be overlooked that the data collection process, processing and subsequent analysis constitute important costs and investments for all the players focusing on products/profit generated by Big Data. An overly long process between the collection of data and their use would nullify their value, which still remains difficult to measure, especially if the data are then sold to third parties.
One of the most recurring problems for companies in the sector is related to the difficulty of satisfactorily connecting data amongst their heterogeneous products (text, images, video, data from social networks, traditional databases...), also succeeding in transforming them into a product that can be validated and consequently used – imagine, for instance, the individual who has to lead budgets/investments on insight and/or analytics with scarce veracity.
The veracity (the possibility that data are not always reliable or even useful) and the complexity of the data to be managed (which is directly proportional to the size of the datasets) are two critical aspects that allow for the Big Data system to be evaluated, not only on the positive outcome of the process, but also on the reliability of the result and the interest that such products will create.
At the end of the process are algorithms that ignite the fuel. Like rough diamonds, Big Data acquire all the facets needed to shine bright in a market that, due to the presence of Big Players, knows how to recognise quality and innovation.
Its benefits are known because more and more, companies are seeing results by applying advanced artificial intelligence algorithms capable of transforming raw data into usable information: NASA chose the Rover’s landing site on Mars by analysing billions of high-definition images of the planet; specialist companies are providing optical tools that automatically recognise skin melanomas...
Automotive: quick to anticipate Big Data!
One of the sectors that has been greatly affected by these changes is the Automotive industry, where the first signs could be seen in 2007 – long before the term Big Data made its appearance.
In a sector where market data are important from the design stage, everything has always been collected, analysed and stored. This applies to vehicles put on the road. This applies to the final users.
Thanks to the availability of information generated by black boxes (acceleration, location, routes and behaviours...) for insurance discounts or by satellite anti-theft alarms, it was realised that a vehicle could become an enormous source of information.
More information flows were born from vehicle-related data. Flows that were used to improve the product, those used to learn customers’ behaviour.
One vehicle can generate an incredible massive flow of detailed data during each second of use: average speed, consumption, performance, reliability, maintenance, driving preferences and styles, accidents, etc.
All of this forms a continuous feed of BIG DATA, ideal for multi-sector monetisation.
In the Automotive industry, this information is important for predictive maintenance (you can improve your product if you know how it behaves), customer care (anti-theft, diagnostics and performance) and infotainment (navigation, traffic information...).
One of the hottest topics is Mobility: once aggregated, the data generated by “connected cars” form the basis for sustainable mobility by improving the journey experience, thanks to the traffic information generated by the very same users on the road. Traffic data are useful, in real time for users, and in the mid- to long-term period for people who deal with smart cities.
The Insurance sector also benefits from the data generated in an automotive context. Not only for Usage-Based Insurance (insurance based on use and driving style), but also in a preventive context (where there’s traffic, there are accidents), civilian context (sensors help to understand accident rate and responsibility) and in the health care sector (driving behaviour generates data that make vehicles, their occupants and transported goods safer).
In terms of the evolution of services inside the vehicle, we must not forget integration with mobile phones. The vehicle’s on-board system will more increasingly become the visual and vocal interface, substituting interaction with the phone.
The car will be a continuation of the current smart phone experience, to increase safety and discourage use when driving, and to have services that we’re now used to having on our phones (messaging, navigation, traffic, park&ride, parking payments...). A new Big Data stream directly from our vehicle.”
To satisfy these evolved demands, however, the market and research no longer concentrate solely on innovative ways to collect more kinds of data, but on new business models to create value around the data themselves. DATA PRODUCTS are created from the analysis of things obtained from the data in terms of “product”, from their integration with other data sources and from their analysis. The main global consultancy agencies have begun to work with Automotive customers to provide data valorisation tools, considered a genuine asset, for the purposes of building business and pricing models.
Only in this way will the market be able to recognise the value and use it.
Thus creating more data to be analysed.