Descriptive and diagnostic analytics

Interestingly, the progression of the four business focused questions we discussed in the previous section — “What happened?”, “Why did it happen?”, “What is likely to happen next?”, and “What are some of the things we can do as a result?” — is the same as the progression of the four key types of data analytics.

The first of these is descriptive analytics

We can answer the first question “What has happened?” through descriptive analytics. Descriptive analytics delivers insights to describe a situation or scenario that happened in the past. It does this by gathering, processing, and summarizing historical data in order to collect useful information and prepare it for further analysis.

A lot of this historical data is produced internally by the organization. The data comes from different departments, such as finance, marketing, distribution, warehousing, etc. We can also run descriptive reports on external data, which is data that has come from any source outside the organization. Common examples of where external data can be found include weather reports, population census, national or federal holidays, geolocation, social media posts, etc.

Descriptive analytics helps reveal patterns and meaning through the comparison of historical data-sets. Insights are not used for making inferences or predictions, which means that we’re unable to answer the question ”Why did this happen?” or “What is going to happen next?” Think of it like a history book that only tells you names, dates and what happened, but doesn’t tell you why these events happened.

The second is diagnostic analytics

Diagnostic analytics describes the techniques you’ll use to answer: “Why did this happen?” It’s a deep-dive into your data to search for valuable insights to identify the root causes of a situation.

Diagnostic analytics adds further context to descriptive analytics by seeking correlations between different data-sets and sets of results. In other words, they’re used to measure or explore the extent to which different things are related to each other, and what effect changing one would have on the other.

Adding this further context requires the data analytics infrastructure to process a lot more data and perform more complex computations than at the descriptive analytics stage. Thanks to the ability for us to capture, clean and transform data using the data analytics infrastructure, and being able to increase the processing power using machines, we’re able to provide quicker, more accurate results to answer “Why did this happen?”

So the difference between the first two types of data analytics is that descriptive analytics is the initial step of data analysis that chronicles the facts of what has already happened. Diagnostic analytics goes a step further to uncover the reasoning behind why it happened.