Traditional data analysis has clear limitations
It could be said that our beginnings in big data and analytics have gone hand in hand. When we started to do analytics in Pragsis we did not have (much) idea, but from 2010 until today we have taken giant steps.
Traditional analytics is an indisputable value when you have the experts with enough business sense and experience to gain a competitive value. However, in the current dynamics of data generation, a traditional analytical approach is somewhat limited.
One of our clients presented us with an opportunity in the industrial sector that represented an important analytical challenge. At that time, the client himself did not expect us to draw much value from our analytical approach. The reality was quite different. After finishing the project, we managed not only to drastically reduce its maintenance cost, but also opened the doors to new analytical possibilities.
Since the founding of Pragsis 15 years ago and especially since our first contact with big data, our technological approach has always been based upon an agnostic point of view. The vision of different solutions and technologies of a changing ecosystem, such as big data is, has had a clear benefit. Not only we manage to offer native analytical solutions in big data, but also we have a privileged vision of the needs and opportunities that clients and projects will need in the immediate future.
Our unconditioned vision in some areas such as analytics, has allowed us to be a company with the capacity to address the most complex needs and be able to be prepared for the current context. This has been partly thanks to our free vision from technical or procedural obstacles, also to the predisposition to which the big data context has prepared us.
Being in this context and assuming analytical challenges, has trained us as analytic natives in big data, and has made us capable of facing needs such as scalability a natural way. The explanation is simple: being experts in big data involves the reengineering of processes and their scaling as core capabilities.
Scalability is not only a challenge, it is also an opportunity and a need. We must be able to move from the limited, frozen and temporary dataset approach to that of an infinitely broader dataset. We must, for example, consider what happens when, instead of having the data for a year, we have all the historical data of a company.
Taking advantage of new opportunities: AI
We must also take advantage of new opportunities like those offered by artificial intelligence (machine learning, deep learning). AI is not a hype and obtaining datasets sufficiently large to feed “the voracious appetite that deep learning has for training data” (Harvard Business Review, 2019). Institutions like the American Association for the Advancement of Science has warned about the risk of the utilization of machine learning techniques derived from limited datasets. (BBC News, 2019)
Data governance as a requisite
But there are still many more challenges. Reasonably, in a context of massive information growth and subject to conditions of legality and privacy, we must also have robust data governance processes.
Final hurdle: productization and support
But all of the above would not do much good if the solution is not capable of productization and support in the organization. These two aspects tend to be the Achilles’ heel of many organizations whose approach to big data has not been realized in a global way.
Again, a broad and global experience in big data and analytics makes Pragsis Bidoop an ideal partner in the exploitation of big data solutions. If it were not true, we would not be here today.
BBC News. (2019). Machine learning ‘causing science crisis’. [online] Available at: https://www.bbc.com/news/science-environment-47267081 [Accessed 6 Mar. 2019].
Harvard Business Review. (2019). Most of AI’s Business Uses Will Be in Two Areas. [online] Available at: https://hbr.org/2018/07/most-of-ais-business-uses-will-be-in-two-areas [Accessed 7 Mar. 2019].