1.
WE LIVE IN AN INCREASINGLY DATA-ORIENTED WORLD: IN BUSINESS PROJECTS AS WELL AS IN THE WORLD OF MEDIA AND INFORMATION, DATA IS OF EVER GREATER IMPORTANCE. HOW HAS THE CULTURE OF DATA CHANGED OVER THE LAST FEW YEARS?
The importance of data, and how powerful it can be, is now widely acknowledged: global players that have built their success on digital phenomena and the power of data (Google, Amazon, Facebook, for example) have clearly demonstrated the business value of data.
Before this was made clear (accompanied by the enabling technology to manage it), data was regarded as interesting as it was useful for measuring known processes in a departmental governance system. It is now clear that analytics based on available data (and not merely on traditionally used data) can give a boost, helping us to consider interconnected systems in a much broader and more sophisticated way, taking the issue of evaluation (rather than that of measurement) to a completely new level.
The challenge now is to move from a data-oriented to a data-driven approach with all due caution and care, as the increasingly popular ML/AI (Machine Learning/Artificial Intelligence) models that can learn from data tend to reproduce the world they photograph, following the changes that are already taking place. For this reason, informed and fully representative data collection is essential for avoiding "bias" in advanced algorithms by influencing their results and even shielding the algorithm itself.
A data-driven approach is certainly very effective in optimisation problems and in improving the performance of complex systems, but is it really useful for innovative leaps? If innovation is to generate discontinuity, there must also be a break in the data that feeds it: a scientific approach consisting of experiments and data verification must be used, but in the most innovative scenarios the data-driven approach must (first and foremost) design a new strategy for collecting useful data.
The importance of data, and how powerful it can be, is now widely acknowledged: global players that have built their success on digital phenomena and the power of data (Google, Amazon, Facebook, for example) have clearly demonstrated the business value of data.
Before this was made clear (accompanied by the enabling technology to manage it), data was regarded as interesting as it was useful for measuring known processes in a departmental governance system. It is now clear that analytics based on available data (and not merely on traditionally used data) can give a boost, helping us to consider interconnected systems in a much broader and more sophisticated way, taking the issue of evaluation (rather than that of measurement) to a completely new level.
The challenge now is to move from a data-oriented to a data-driven approach with all due caution and care, as the increasingly popular ML/AI (Machine Learning/Artificial Intelligence) models that can learn from data tend to reproduce the world they photograph, following the changes that are already taking place. For this reason, informed and fully representative data collection is essential for avoiding "bias" in advanced algorithms by influencing their results and even shielding the algorithm itself.
A data-driven approach is certainly very effective in optimisation problems and in improving the performance of complex systems, but is it really useful for innovative leaps? If innovation is to generate discontinuity, there must also be a break in the data that feeds it: a scientific approach consisting of experiments and data verification must be used, but in the most innovative scenarios the data-driven approach must (first and foremost) design a new strategy for collecting useful data.
2.
THE FIGHT AGAINST COVID-19 HAS DEMONSTRATED THE IMPORTANCE OF OPEN DATA AND, IN GENERAL, OF DATA INTEROPERABILITY: IN A PROCESS OF DIGITAL TRANSFORMATION AND IN THE CREATION OF DIGITAL ECOSYSTEMS, HOW IMPORTANT IS "INTEROPERABLE" DATA THAT CAN COMMUNICATE WITH DIFFERENT SYSTEMS
The interoperability of data is fundamental. The limit of what can be done with data is dictated by the limits of the data itself. To be interoperable, the data must be:
Open data represents a great resource, but it tends to reach only a minimum of its expected potential: it is often published at such levels of aggregation as to be effectively useless, without continuity in time and space, or more often as isolated snapshots that are unusable for in-depth and broader studies of general phenomena.
Interoperability as a technical problem is easily overcome; but the real prerequisite is the stable and representative availability of extensive high-quality data. For this reason, in a process of Digital Transformation, we often think about how to start producing useful data and not just what to do with the available data.
The interoperability of data is fundamental. The limit of what can be done with data is dictated by the limits of the data itself. To be interoperable, the data must be:
- available with adequate depth in time and space
- understandable, with explicit and clear semantics
- guaranteed over time.
Open data represents a great resource, but it tends to reach only a minimum of its expected potential: it is often published at such levels of aggregation as to be effectively useless, without continuity in time and space, or more often as isolated snapshots that are unusable for in-depth and broader studies of general phenomena.
Interoperability as a technical problem is easily overcome; but the real prerequisite is the stable and representative availability of extensive high-quality data. For this reason, in a process of Digital Transformation, we often think about how to start producing useful data and not just what to do with the available data.
3.
WHAT IS THE IMPORTANCE OF DATA VISUALISATION IN OPTIMISING THE VALUE OF DATA? WHICH IS THE RIGHT APPROACH AND WHICH ERRORS MUST BE AVOIDED?
Data visualisation is merely the tip of the iceberg, but in many cases it is what allows all the work underneath to emerge and shine. Data visualisation is to the world of data analysis what user-experience is to the world of apps: it marks the difference between a successful app and one that is unused, even if it carries a much higher intrinsic value.
Vital and interesting projects (including the most advanced ML/AI algorithms) run the risk of being thwarted and remaining merely exercises in style if the results (as valid as they may be) remain expressed in the typical formats used by experts and if efforts are not made to render them comprehensible to non-expert users and make them accessible to business users. Data visualisation is useful in precisely such contexts, helping to make traditional information and the results of complex algorithms easy to use and comprehensible to a target audience of users that are focussed on the business and on meanings, unlike their producers.
To be effective, data visualisation follows representation paradigms linked to our perceptual capacities as well as communicative paradigms that promote the immediate understanding of relevant aspects. In this context, excellence is achieved by creating almost tailor-made graphics and pursuing the best experience for the benefit of target users and analysis objectives.
Data visualisation is merely the tip of the iceberg, but in many cases it is what allows all the work underneath to emerge and shine. Data visualisation is to the world of data analysis what user-experience is to the world of apps: it marks the difference between a successful app and one that is unused, even if it carries a much higher intrinsic value.
Vital and interesting projects (including the most advanced ML/AI algorithms) run the risk of being thwarted and remaining merely exercises in style if the results (as valid as they may be) remain expressed in the typical formats used by experts and if efforts are not made to render them comprehensible to non-expert users and make them accessible to business users. Data visualisation is useful in precisely such contexts, helping to make traditional information and the results of complex algorithms easy to use and comprehensible to a target audience of users that are focussed on the business and on meanings, unlike their producers.
To be effective, data visualisation follows representation paradigms linked to our perceptual capacities as well as communicative paradigms that promote the immediate understanding of relevant aspects. In this context, excellence is achieved by creating almost tailor-made graphics and pursuing the best experience for the benefit of target users and analysis objectives.
4.
IN THIS CONTEXT, HOW HAS OUR CUSTOMERS' APPROACH TO BUSINESS INTELLIGENCE CHANGED?
Having overcome the obvious demands for mobility and self-service in BI solutions, what is now being sought is pervasiveness, i.e. the ability to provide data where it is needed, when it is needed, and to integrate it seamlessly into different applications.
Individuals are used to using simple and intuitive smartphone apps, which provide large volumes of information at the click of a button; this now undermines the arguments of difficulty and delay in making company data available. Rigidly structured approaches no longer work or are insufficient: they need to integrate a parallel high-speed line (see also the bi-modal approach recommended by Gartner), which allows fast responses to new demands while operating in a high-quality and robust way on more traditional pathways.
In order to satisfy this twin-track approach, today we can also make hybrid choices, both in terms of data architecture (data lake, dwh, on-prem, cloud, hybrid-cloud) and application services. The conventional route, however, must not be completely abandoned, otherwise we risk falling into the paradox of Excel (as utilised by the average user), whose strength lies in its limits: it is easy to use, completely free (users can do what they wish with their data), and offers quick results - but these are always open to question because they depend solely on the practices of the individual user.
Business Intelligence must provide freedom and speed in meeting new demands, but it must always guarantee a single and shared form of data semantics as a prerequisite for the interoperability mentioned above.
Having overcome the obvious demands for mobility and self-service in BI solutions, what is now being sought is pervasiveness, i.e. the ability to provide data where it is needed, when it is needed, and to integrate it seamlessly into different applications.
Individuals are used to using simple and intuitive smartphone apps, which provide large volumes of information at the click of a button; this now undermines the arguments of difficulty and delay in making company data available. Rigidly structured approaches no longer work or are insufficient: they need to integrate a parallel high-speed line (see also the bi-modal approach recommended by Gartner), which allows fast responses to new demands while operating in a high-quality and robust way on more traditional pathways.
In order to satisfy this twin-track approach, today we can also make hybrid choices, both in terms of data architecture (data lake, dwh, on-prem, cloud, hybrid-cloud) and application services. The conventional route, however, must not be completely abandoned, otherwise we risk falling into the paradox of Excel (as utilised by the average user), whose strength lies in its limits: it is easy to use, completely free (users can do what they wish with their data), and offers quick results - but these are always open to question because they depend solely on the practices of the individual user.
Business Intelligence must provide freedom and speed in meeting new demands, but it must always guarantee a single and shared form of data semantics as a prerequisite for the interoperability mentioned above.
5.
WHAT IS THE VALUE OF A SUITE LIKE KNOWAGE AND HOW DOES IT CONTRIBUTE TO A COMPREHENSIVE AND INTEGRATED USE OF DATA?
KNOWAGE was born and grew as an open and flexible solution, in the open-source spirit of sharing and agility. One of its strengths lies in the fact that is allows the combination of data from different sources (including sources that are technologically very different), thus enabling the hybridisation of data, sources and storage technologies, as well as the coexistence of complementary methodological approaches, with the possibility of managing the double speed mentioned above.
For data visualisation it offers a wide range of predefined graphics; above all, however, it allows the insertion of customised graphics and total customisation of the visual experience, even allowing tiered infographics if necessary.
In order to meet the needs of increasingly widespread complex analytics, it allows the integration of R/python functions, bridging the gap between data science and data visualisation for that last mile, which can enhance its effectiveness.
In addition, KNOWAGE allows the creation of interactive viewpoints with the insertion of navigation paths between the data that supports the user in the process of analysis, making it possible to zoom in, focus on single assets, and translate points of view, promoting freedom of thought in the process of analysis, which should - when presented with a piece of clear information - always generate further questions.
KNOWAGE was born and grew as an open and flexible solution, in the open-source spirit of sharing and agility. One of its strengths lies in the fact that is allows the combination of data from different sources (including sources that are technologically very different), thus enabling the hybridisation of data, sources and storage technologies, as well as the coexistence of complementary methodological approaches, with the possibility of managing the double speed mentioned above.
For data visualisation it offers a wide range of predefined graphics; above all, however, it allows the insertion of customised graphics and total customisation of the visual experience, even allowing tiered infographics if necessary.
In order to meet the needs of increasingly widespread complex analytics, it allows the integration of R/python functions, bridging the gap between data science and data visualisation for that last mile, which can enhance its effectiveness.
In addition, KNOWAGE allows the creation of interactive viewpoints with the insertion of navigation paths between the data that supports the user in the process of analysis, making it possible to zoom in, focus on single assets, and translate points of view, promoting freedom of thought in the process of analysis, which should - when presented with a piece of clear information - always generate further questions.
Grazia Cazzin
Grazia is the Offering Manager of the Data & Analytics competence center in the Research and Innovation Division of Engineering Group and the KNOWAGE Labs Director.
Working in the IT field since 1992, she has been involved in enterprise application development, data modelling, data warehousing, dimensional analysis and business intelligence.
Currently she is involved in all the activities related to modern data architectures, advanced analytics and data visualization, as in Augmented City, Smart Agriculture, E-Health and Digital Industry.
Working in the IT field since 1992, she has been involved in enterprise application development, data modelling, data warehousing, dimensional analysis and business intelligence.
Currently she is involved in all the activities related to modern data architectures, advanced analytics and data visualization, as in Augmented City, Smart Agriculture, E-Health and Digital Industry.
Contact us
marketing@eng.it
Tel. (+39) 06-87591