So, let's take a minute to look over what we just did:
- We received a file of data that contained the results of patient dialogs, along with some pertinent patient history.
- We looked at some of the data and then went ahead and loaded it into Watson Analytics.
- Anxious to create a prediction of our data, we started by creating a New Prediction, but rather than completing it, we used the Field Properties feature to scroll through the columns of data so that we could obtain an understanding of what we had to work with.
- At this point, we determined that we didn't know enough to move forward, so we used the IBM Watson Analytics Explore feature to dive deeper into the data.
- From there, we were able to build upon Watson's prompts and establish a target for our prediction of Number of Illnesses. We also found some interesting insights into the data, such as the fact that the use of recreational drugs as well as smoking seems to not materially affect the number of illnesses for a patient, but the use of alcohol does.
- We were able to glean these insights from the data quickly via Watson Analytics's generated visualizations. Since these seem key, we saved each of these visualizations as assets to our collection.
- At this point, we know our data pretty well, and we have some solid ideas in mind of what might drive or affect our prediction, so now it's time to create our prediction.
- If we click on the file's tile from the Watson Analytics Welcome page, Watson again gives us some starting points. Since we think we have a good idea as to how to proceed with our prediction, we can just click on the Predict icon at the bottom of the dialog, as shown in the following screenshot:
- Watson will then present the Create a new analysis page, as shown in the following screenshot:
- We can fill in a name and select the target that we are interested in (such as Number of Illnesses), as shown in the following screenshot:
- Finally, we can click on the button labeled Create, and Watson Analytics will begin its analysis of our data, as shown in the following screenshot:
