Today predictive analytics is a process that consists of three distinct but interrelated parts:
Capture refers to capturing data.
But what do we mean by data? Let us use the data triangle methodology to understand data and its various aspects.
- At one corner of the data triangle we have sources of data. These could include traditional sources such as:
- relational databases
- flat files
- Excel spreadsheets
- Sources could also mean big data sources like Hadoop, SQL systems and other analytic data stores.
- On the second corner of the data triangle we have forms of data.
This would include data at rest which refers to data about transactions that have occurred in the past and data in motion, also sometimes referred to as streaming data.
- On the third corner of the data triangle we have types of data.
This includes structured data which is data stored in rows and columns and unstructured data which includes freeform text pictures, videos and so on.
Second part of the predictive analytics process: predict.
Once you have the data the next step is to make predictions off the data by using various techniques including:
- data mining;
- text mining;
- statistical analysis.
And finally after making predictions you must act upon those predictions.
Not acting upon insight is like jumping off an airplane without deploying the parachute strapped to your back: it is pointless and cause unintended results.
Let’s recap: predictive analytics is a process where the first part is capturing data which can be discussed using the data triangle methodology. The next step is predict, which is done using techniques like data mining, text mining and statistical analysis.
The last step of the predictive analytics process is “To Act” which involves acting upon the insight from the predict stage.
Learn more how Metricus uses predictive analytics to enable ITSM professionals. Visit Metricus Platform Features page.