top of page
image.png

Big Data: "A source of growth for every industry"

  • sarah61533
  • May 26
  • 3 min read

Thanks to the development of specific techniques, Big Data has benefited many different sectors of the economy through its ability to process large amounts of information. From boosting business performance to new governance requirements, we look at how Big Data has changed the way companies handle information. Ugo Comignani, a lecturer at Grenoble’s ENSIMAG school of IT and applied mathematics and a co-director of a Masters program specialising in Big Data run by ENSIMAG and the Grenoble School of Management, is our guide.


Big Data: A driver of performance


Faced with a gradual increase in the amount of data being generated around the world, analyst Douglas Laney coined the phrase 'Big Data' back in 2001. Describing its main characteristics in terms of the '3 Vs' -- volume, variety and velocity – it meant that such data existed in vast volumes, had various types (structured, semi-structured or unstructured), and was produced, stored and processed at ever-faster speeds. Irrespective of their industry, it has proved to be revolutionary for many companies.


As Ugo Comignani explains: "The growing quantity of data and the development of techniques to analyse it has led many companies to reconsider the importance of data in their activities. Today, it’s no longer seen as an ‘extra’, but as an essential asset for managing the business and generating value.” By rationalising decision-making, improving customer knowledge, detecting errors or fraud, identifying new business opportunities...“data is a source of growth for all industries,” says the expert. "And Big Data makes it possible to exploit that growth even more effectively.”


Big Data certainly has many applications in the property sector. Algorithms can be used to cross-reference data, for example, to match people offering properties for rent with potential tenants, or to compare different properties in order to assess their likely appeal to buyers. Before investing in a property, it can also be very useful to combine information provided by the vendor with other types of information, such as the local weather. This type of open data can inform potential investors about the presence of any natural risks in the surrounding area the property in question. "By aggregating all this information, you get a much more accurate view of the risk-reward ratio, as it’s based on a larger volume of data. The estimate of a property's likely profitability will therefore be much more fact-based, which in turn will make the investment decision a much more pragmatic one," summarises Ugo Comignani.


Data quality, model quality


As a result, Big Data has revolutionised the way companies manage their businesses and has profoundly changed the way they are organised, notably with the recruitment of technical experts and the formation of departments dedicated to data. As Ugo Comignani points out: "The increase in data flows has created a major challenge for companies -- the governance of all this data. Poor quality data cannot be used, but it’s very difficult to identify once it has entered the system. To avoid this pitfall, you need specific codes, roles, and a rational structure for a company’s data.


Furthermore, he insists that good quality raw data is no guarantee of good quality analysis: the real key is to build the right models for analysing that data. "If your data is excellent but your machine learning model isn’t suitable, you won't get very far," he warns. So, although better quality data can lead to better datasets for training the models, those datasets need to be truly representative of the population being studied, the algorithm must be adapted to the problem being tackled, the learning phase needs to be sufficiently complete, etc. If these conditions cannot be met by data scientists, the risks include an ‘over-fitting’, i.e. building an algorithm that is only really suitable for a single dataset, a situation which is often the result of a lack of training with a sufficient amount of varied data.


Data storage


Meanwhile, the amount of data being generated continues to grow. For Ugo Comignani, "the accumulation of data creates new problems, both in terms of the technical capacity to record and store this data and the cost of these new tools ". The volume of data produced in 2020 is estimated to have been around 60 exabytes. However, between 2020 and 2021, only 2% of the data produced will have been stored, according to estimates. "We are now producing much more data than we can store. One of the main future challenges for Big Data is the storage resources, along with the ethical treatment of this data, which is often of a sensitive nature," he concludes. Data quality, storage capacity and the ethical treatment of information: Big Data has raised a number of issues about the corporate governance of data, some of which are still in their infancy.

bottom of page