The importance of analytics

Everybody is talking about analytics. Together with Artificial Intelligence or AI all of our business problems will be solved apparently.

Analytics sounds like analysis, so it is natural to make the comparison to try and understand what is different between the two.

Analysis is defined as “the process of breaking a complex topic or substance into smaller parts in order to gain a better understanding of it.”. In common parlance, analysis often means the act of using  quantitative statistics to explain or discover something of interest.

Analytics is explained as “the discovery, interpretation, and communication of meaningful patterns in data. It also entails applying data patterns towards effective decision making.”.

At first glance, there isn’t much of a difference between these two statements, which I am sure does not help people understand any distinction between the two terms.

For me, analysis remains the core activity of reducing the complexity of data so that it can be comprehended. Analytics is much broader than this, as not only does it include the methods and tools required to create a platform for analysis to take place, it also includes the context in which the data to be analysed resides. Analytics is the scientific thinking and processes behind analysis, the whys and wherefores, and therefore analysis is a component of analytics.

In the manufacturing domain at least, predictive analytics is very topical, and in the broader business domain in general, decision analytics is popular with enterprise software vendors.

Predictive analytics is essentially forecasting, which itself is a mature statistical subject. It is a human trait to want to understand and plan for the future, and considerable research and experience has developed knowledge in this field.

Being able to predict behaviour using a set of input variables enables transport operators to replace service items more economically than using traditional planned maintenance schemes. Similarly, while two different items of plant may both utilise the same bearing type, the difference in work loading on each machine may result in different wear patterns and therefore service life.

It is therefore more prudent to replace either bearing when it is predicted to approach its failure, rather than on a pre-determined date. Such a scenario is described as predictive maintenance, and often includes the topic of condition monitoring.

Businesses often want to identify segments in their customer base, in order to develop ideas for innovative products and services that might appeal to those customer types. It requires analysis that can take a collection of data and identify the characteristics that enable that data to be classified into discrete groups. This is referred to as decision analytics.

Taken together, both predictive analytics and decision analytics are inherent parts of digital manufacturing, and are thus commonly referred to when discussing Industry 4.0.

Inexpensive hardware is assisting the adoption of Industrial Internet of Things, and this is creating a deluge of data that needs to be analysed and visualised in ways that aid its comprehension.

Analytics thus helps us not only understand the insight that lies within data, but it also assists how we cope with increased data volume by way of providing the tools, platforms and methods to manage and analyse that data.