To gain a better understanding of how to use the Watson Analytics Predict and Assemble features, let's now take a look at a simple use case.

One of the best ways to learn a new tool is by using it, and to use Watson Analytics, you need data. Up to this point, we've utilized sample data for use cases that I created from various sources, but Watson Analytics has made many sample datasets available for use for your learning. To view the sample data options, simply click on Add from the main or Welcome page and then click on Sample Data, as shown in the following screenshot:

For our use case, we'll go back to an example used earlier in this book—Stadium Sales. For this use case, we've received a similarly formatted file, but one that includes historical results of products sold by a particular NFL team at their home stadium over two previous seasons. The file, named Historic_Stadium_Sales, can be uploaded to Watson Analytics as explained in earlier chapters:

Once the file is uploaded (shown in the previous screenshot), you can click on the upper portion of the file's tile and then select Predict:

Remember, that you could've used the Watson Analytics Refine or Explore features on this file (as we have already covered), but since this file is really the same as the original stadium sales file, we feel that we are relatively comfortable with its format, so we'll just go ahead and try Watson Analytics' Predict feature.

After clicking on Predict, let's name our workbook (I've decided to call it Historic Sales) and set a target (I picked Product):

Next, click on Create. Watson Analytics, in real time, runs its analytical algorithms on our data and displays the following insights:

Watson Analytics has organized its results into various sections, so we can understand them more easily. Let's start from the center of the page, as shown in the following screenshot:

This section is focused on the predictors of our selected target (Product). In other words, which fields in our data have some influence on the value of Product? This is interesting, since I may be trying to determine which particular product sells best and when. Watson Analytics has found that three fields have some correlation to Product: Winning team, Week, and Season. I can easily set that by sliding my mouse arrow over the three bullets in the target:

For example (as shown in the preceding screenshot), Winning Team is displayed as a predictor bullet of Product (as well as the other fields, Week and Season). From the bullet popup, you can see that Predictive Strength of Winning Team is over 27 percent, and you have the option to click on two links: More details… and Associated fields….

To the right of our target, the predictors are displayed in another way: What influences Product? Check out this screenshot:

Since the top predictor (and the most intriguing one to me) is the Winning team field, let's look a bit closer at it. If you click on the More details… link from the bullet popup, or on the top visualization to the right of the target, Watson Analytics zooms in to the Main Insight:

We can easily see from the visualization provided that when the home team wins, the most popular product seems to be Team Hat - Cap. This is easy for me to understand, but if you are interested in the statistical details of this insight, you can click on Details in the top-right corner of the visualization (shown here):

Since you are so inclined, Watson Analytics provides a brief explanation of how it arrived at the main insight (winning team drives product). Across the top of the details page, you see this statement:

Product is a categorical target, so a logistic regression based approach is used.

If you look closely, you'll notice that categorical, target and logistic regression are hyperlinks that will provide definitions and Learn more links if you click on them, as marked here:

All throughout the prediction, look for the presence of these helpful hyperlinks to build your knowledge of the statistical theory behind Watson Analytics' insights. In addition, wherever possible, click on the Collect icon (shown in the following screenshot) to add to your collection of artifacts to be used later:

Back to the Top Predictors section, to the bottom left, we see More Predictive and Easier to Understand:

Watson Analytics starts with the easiest—One Field selected as a start. The One Field is Winning Team. You can experiment by switching to Two Fields (what two fields drive the product?) or Combination (is there a combination of fields that drive the product?) and see the results of your selection in real time.

Across the top of the prediction is the basic information band, as shown in this screenshot:

Here, Watson Analytics provides basic information, such as the following:

Across the bottom of the page is Data Tray; it lists all of the fields within our data file, as shown here:

From the Data Tray, you have the ability to drag fields into your prediction. For example, you can select Quantity and drop it onto the target to see that fields Top Associations. As shown in the following screenshot, Watson Analytics tells us that the Winning Team field is also associated with the value of Quantity; that is, when the home team wins, we have a higher number of products sold:

Finally, in the top-left corner is the MENU icon, from where you can access important Watson Analytics features, that is, FIELD PROPERTIES and VERSIONS:

FIELD PROPERTIES lets you explore and change the statistical properties of each field in your data file. The fields are listed vertically on the left and the properties of the (selected) field are on the right. You can sort and filter the fields if you need to (if the number of fields is large, this is very helpful):