Analytics is a vast topic, and sometimes even those who know some of the relevant techniques can get lost when reading up on something new in the field. Part of this confusion can stem from the fact that there are different types of analytics, and while there is some overlap between these types, what they are actually used for can be quite different. The first of the three covered in this series is descriptive analytics.
Descriptive analytics, as the name suggests, describe things. Specifically, they take information from one or more sources and use it to describe things that have already happened, allowing the past or present to be interpreted more easily. These analytics then allow us to learn from what has happened already, so we know our current position better.
What does it look like?
Think about the dashboard of a car. You have the speedo, rev-meter, engine temperature and fuel gauge. Some fancier models might even have a meter displaying a measure of how efficiently you’ve been driving during your journey. These are all descriptive analytics.
“Without the proper context, descriptive #analytics may actually mean very little.”
The speedo will be taking information about how fast the wheels are spinning and showing you how fast you’re going without you having to count the revolutions of the wheels yourself. Similarly, the journey efficiency meter will take fuel level measurements and recordings from the rev-meter and, using an algorithm, will provide a single number that summarises the efficiency of your past driving throughout the journey up to that point.
Importantly, the measures above may potentially be really useful, but they don’t mean much without context. The speedo might read 50mph, but without knowing the speed limit this information would be meaningless.
Descriptive analytics are probably the most common analytics used in practice and reporting. Many terms will be familiar, but as a quick reference here are a few of the more obscure ones.
|Demographic||Descriptive analytics will often report common customer attributes, which may fit into a certain category of customer. This is known as demographic information.|
|“As at <date>”||Often used in financial reporting, this doesn’t really sound like fluent English and isn’t a descriptive analytic. All it means is that the descriptive analytics in the report were correct up to the date provided.|
|Metric||A measurement. This may be an established metric like kilometres, grams or some other representative value.|
|Mean||This isn’t someone being nasty. Instead, it is what most people know as the average. Add up all the values and divide by the total number of values added.|