Analytics Scrabble - and how to make sense of it

Data Science, Advanced Analytics, Machine Learning, Predictive Analytics, Statistics, Data Engineering, Business Intelligence, Data Mining, Business Analytics…and the list continues...Actuarial Sciences, Prescriptive Analytics, Big Data...getting confused what these really mean?

I started my IT consulting career with database and reporting development long time ago and in the latest jobs have been leading sales in the most advanced analytics areas, however even with decent background it is sometimes bit time consuming to understand the essence of the continuously evolving vocabulary. For me putting the terms on a chart with dimensions that reflect the needed capabilities and skills has helped a lot in positioning the relations of the terms. The matrix is far from complete and actually quite trivial, but hopefully also helps others in gaining further understanding of the terms.

Starting from the IT side, the first term is Data Engineering, which is about consolidation of data and passing that in somewhat cleansed format for use for Data and Business Analysts or Data Scientists. This is very much an IT focused discipline.

Next one is Data Analysis, where the data is analysed from narrow, rather technical and statistical point-of-view. Business Intelligence is an interesting animal. Consolidated data from the Data Engineers are consumed by business professionals that have basic analytics understanding but quite deep business understanding. As there is really no focus in true data mining or algorithm development these are often called Business Analysts. One could actually label them as “light-weight” data scientists.

Now we start to enter the Data Science domain. Add bit more business insight into Data Analysis and you get into Business Analytics, which might focus in very specific industries. E.g., actuaries working with insurance premiums and their prediction models fall into this category. The step that brings us into the core of Data Science and when we start to talk about Data Scientists is when the activities are circled around data mining and algorithm development. Interestingly, Gartner has renamed Advanced Analytics to Data Science mostly since it is Data Scientists working in this area.

In Data Science, Machine Learning computational capabilities are utilized to develop algorithms and models for predictive and prescriptive analytics. Predictive analysis is about predictions what will happen while prescriptive analytics is about recommendations of actions. So what’s then Big Data? We talk about large amounts of unstructured data that the Data Scientists will utilize in their Data Mining activities to extract insight. This could also be called Knowledge Discovery, but we are essentially talking about the same thing.

Hope the structure I briefly walked thru will help you to gain understanding of some of the Analytics jargon. This is good starting point but there is lot more to cover e.g. I did not touch how Artificial Intelligence is related to these. I will come back to that point later. Also, when walking through some of the analytics learning materials one bit disturbing note came across about deployment of models or algorithms developed in the Data Science Process: “Once the proof of concept is finalized, 95% of projects stop there”. Sounds like expensive experiments, but will come back on that as well.

Stay tuned, Petri