Howard Altmann
Mar 23, 20212 min
Analytics is the science of analyzing raw data in order to achieve a better
understanding of the raw data and/or system and to make conclusions about that
information. The aim of analytics has been said to provide better support to the
decision-making process.
There are four types of analytics – descriptive, diagnostic, predicative and
prescriptive.
Descriptive analytics consider the historical aspect of the data. It can organize the
data into a convenient form including, but not limited to graphs, tables and control
charts to further understanding for non-analysts. It is said to answer the question of
what happened.
Examples of descriptive analytics include:
Summarization of a particular business result after a change is made to a
process
Determination of whether a quality problem exists in the manufacture of a
product
Diagnostic analytics consider identification of positive and negative anomalies that
are uncovered after looking at the output of the descriptive analytics and trying to
determine why that (those) anomalies occurred. Diagnostic analytics is synonymous
with root cause analysis.
Examples of diagnostic analytics include:
Determination of the cause of a spike in sales of a product, or better yet,
sustained increase in sales
Determination of why a product unexpectedly deviates out of its control
parameters and worse yet, its specification parameters.
Predictive analytics utilize data and statistical models to determine what might
happen with various data scientist given scenarios. It is complex and uses
sophisticated tools. Wikipedia writes that the following scientific disciplines
comprise predictive analytics
Machine Learning
Operations Research
Computer Vision
Natural Language Processing
Applied Statistics
Signal Processing
Image Processing
Metaheuristics
Date Mining
Examples of predictive analytics include:
Using spending patterns, outstanding credit balances, income sources and other
financial measures to calculate credit scores which will predict a person’s ability
to afford a mortgage, a new credit card or an auto loan.
Using spending patterns to identify excess spending such as might occur when
one’s credit cards are stolen.
Prescriptive analytics suggests a range of prescribed actions and the potential
outcomes of each of the actions. It may even go so far as recommending one of the
prescribed action / potential outcomes as being superior.
Examples of prescriptive analytics include:
Self-driving cars. Given a location and destination, prescriptive analytics will
consider a range of driving paths, speeds and traffic to safely travel from start to
finish in the shortest amount of time. It also updates the prescription as time
goes on.
Air travel. If bad weather suddenly appears at an airport, prescriptive analytics
will consider how to safely hold and/or re-route incoming traffic until conditions
improve and route these to the air traffic controller.