Descriptive, predictive, prescriptive: types of analytics

In a last entry of the blog, we told you what Data Analytics was. Today we are going to tell you a little more about what types of data analytics there are and the applications of them. Before that, it is necessary to talk a little about the kinds of data that exist. As you can see in the image, some data are structured, semi-structured and others are unstructured. Structured data are those stored in your excel sheets or a database, and are characterized by having a predetermined order; we almost always organize the data in rows and columns. On the other hand, unstructured data does not have a predefined form, for example, the photos you post on Instagram, the videos you record on your smartphone, the voice messages that you send via Whatsapp, what you post on your Facebook wall, among others. Semi-structured data have something of both types.

Now, as we said, there are several types of analytics. The level that can be reached within an organization depends on the maturity of it. We will see the three best known and future entries will give more detail of the other types and what is that of business intelligence).

Analytics subfields
Source: Rapidminer (s.f.)

Descriptive Analytics

Generally, one begins by doing descriptive analytics. So what is that? Think about when your boss says “I want the descriptive statistics (frequency tables, histograms, bar graphs, average, mode, median) of that data.” Yes sir or madam, there you started to apply data analytics. Of course, descriptive analytics goes a little beyond that. As the name implies, this type of analytics intends to describe in the best possible way a situation, event or product to answer what happened.

Predictive Analytics

According to Kelleher Mac Namee and D’Arcy (2015), predictive analytics consists of the art of constructing and using models to make predictions based on the patterns extracted from the historical data of interest. Now we are going to answer the question that inevitably arose: And this kind of analytics, what is it for?

This type of analytics answers what it is most likely to happen. We do not enough space for describing everything we can do with this type of analytics, so in this post, we will tell you only about two uses we can give it take and complement those describe by Kelleher Mac Namee and D’Arcy (2015):

It is essential to keep in mind that, for any of these applications, in addition to having a human resource or an expert provider in analytical solutions, you need a historical database. Moreover, that does not mean a year or two of data, that means a minimum of more than three years if we are talking about a monthly periodicity and remember that the more periods you have, the better the accuracy of the model, which will never be 100%. If you do not have the data, but you recognize that your organization has problems that could be solved or opportunities that could be exploited using predictive analytics, then it is time to start thinking strategically about the capture and storage of that data that you need. histórica. Y eso no significa un año o dos de datos, eso significa mínimo más de tres años si estamos hablando de una periodicidad mensual y recuerde que entre más periodos tenga mejor será la precisión que podrá tener el modelo, la cual nunca será del 100%. Si no tiene los datos, pero usted reconoce que su organización presenta problemas que se podrían solventar u oportunidades que se podrían aprovechar utilizando la analítica predictiva, entonces es hora de empezar a pensar de manera estratégica sobre la captura y almacenamiento de esos datos que usted necesita.

Prescriptive Analytics

Prescriptive analytics goes a step further. According to Bull, Centurion, Kearns, Kelso, and Viswanathan (2015), the interest for it boomed in 2013 (I have not yet found the reason why this is so) although the methods that are used exist a long time ago. Prescriptive analytics is supported by techniques of operations research, machine learning, applied statistics and natural language processing to answer questions such as: machine learning, estadística aplicada y procesamiento del lenguaje natural para responder interrogantes como:

  • How to optimize the trading strategy?
  • How to optimize the financial services portafolio of my company?
  • How to optimize the mix of products offered by my company?
  • What is the best or the worst that can happen?


  1. Bull, P., Centurion, C., Kearns, S., Kelso, E., y Viswanathan, N. (2015). Prescriptive analytics for business leader. River Logic.
  2. Chartered Global Management Accountant. (2016). Business analytics and decision making: The human dimension.
  3. Kelleher, J. D., Mac Namee, B., & D’Arcy, A. (2015). Fundamentals of machine learning for predictive data analytics: algorithms, worked examples, and case studies. MIT Press.
  4. Rapidminer. (s.f). An Introduction to Advanced Analytics. Disponible en:

Leave a comment