Analytical Journey

In the world of data analytics there are many different analytical methods available. Understanding what can be done with the different methods can be challenging and confusing. An analytics journey within an organization typically starts with with historical reporting, with long term aspirations of moving into machine learning and artificial intelligence.

Between the two endpoints, reporting and artificial intelligence, are a number of different options. Moving along the analytics journey the methods increase in complexity.

Basic	Intermediate		Advance
Reporting	Inference	Probability	Clustering	Prediction	Prescriptive	Artifical Intelligence
What happened	Is what happened significant	Chances of something happening	What is similar	What could happen	How to take action	Continually evolving

What Happened

The use of this method is retrospective, providing insight on things that have happened. Examples of what happened are:

• How many products were sold last month
• What is the average number of visitors on a Friday
• What is the minimum amount of time spent visiting a site

Is What Happened Significant

The use of this method is to make estimates of a population, or drawing conclusion of a hypothesis. Examples of is what happened significant are:

• Was yesterdays usage normal
• Is the increase in energy usage an outlier
• Was the amount of precipitation last month an anomaly

Chances of Something Happening

The use of this method is to estimate the likelihood of an event happening. Examples of chances of something happening are:

• If a weather event will cause a travel delay
• Possibility of a customer making a return
• Likelihood of client using a new service

What is Similar

The use of this method is to identify groups that share similar properties. Examples of what is similar are:

• Create customer profile segmentation
• Classify travel delays
• Identify hotspots where health outbreaks occurred

What Could Happen

The use of this method is to forecast potential future outcomes. Examples of what could happen are:

• Estimate wait time under different conditions
• Forecast new service usage
• Scenario planning

How to Take Action

The use of this method is to leverage past outcomes and trends to identify what needs to completed to achieve future goals. Examples of how to take action are:

• Recommend optimal service plan
• Prioritizing work orders with the least operational impact
• Selection of products with highest revenue potential

Continually Evolving

The use of this method is to identify unseen outcomes. Examples of continually evolving are:

• Customer chatbot
• Demand based pricing
• Future based evaluation of work schedules