Monday, September 26, 2016

Data Science - Type of Analytics

While discussing with few data enthusiasts, the topic Type of Data Analytics came up and various types were discussed. There were few arguments on categorization but it is clear that the types of Data Analytics are vary and can be defined with many classifications.

However, if someone needs to understand main types of it, what would be the way of categorizing it. As per the discussion we had and the articles I have read, I believe that it can be divided into three;

If we are to summarize this;

Descriptive Analytics is something we always do, we always implement. We try to see the insight based on historical data. You might be analyzing last 6 months supermarket bills and try understand how you have spent, whether you have spent on unnecessary items, whether your spent is worth or not, etc. The owner of the supermarket would do the same. He will check and see how his customers have bought items, any relationships among purchased items, which gives more profits, which do not move, etc. Generally this is the first part of any Business Intelligence implementation and this is what widely used by many users.

Predictive Analytics helps us to see the future. This is also based on historical data but it actually forecasts values for future. This analytics uses historical data along with statistics related to all entities. This is not only for seeing values, it can be used for other processes related to the business. Example, we can use this analytics for finding associations between products and keep those products closer to increase the sale. That is why we see some different names on this such as Market Basket Analytics, Association.

Prescriptive Analytics combination of both above items but it shows few things additionally. It tells you what might happen and what should be done. This analytics is bit rare to see with standard BI applications but many try to implement it. This is referred as Describe - Predict - Prescribe

No comments: