Field roles determine how a particular field is used in a prediction. The role of a field can be changed by selecting the desired value from the Role drop-down list, as you can see here:
Measurement Level is used to classify a field for predictive analysis and can affect Watson Analytics' ability to identify insights. The measurement levels for a field can be changed by selecting them from the Measurement Level drop-down list, as shown in this image:
Data issues are also visualized here. In our example, the Gameday Weather field is supposed to hold the weather conditions on the day of the game. In our file, Watson Analytics has found that all records contain Sunny, which appears suspicious, so it excludes the field from further analysis:
When you change anything in the field properties, Watson Analytics automatically creates a new version of your prediction for you. Under Menu and then Version, you can view and access/reload any of those saved versions, as follows:
You should now understand Predict—at least enough to start experimenting with your data, so let's move on and take a look at Assemble.
As we've mentioned earlier in this book, Assemble gives you the ability to present what you may have discovered using Watson Analytics in various ways, such as dashboards and stories. To illustrate, let's assemble a dashboard using our historical stadium sales prediction example that we just walked through.
From the Watson Analytics Welcome page, we click on Add and then on Assemble, like this:
From the Create new View dialog, we select (double click on) our data, Historic_Stadium_Sales:
Watson Analytics will then display the Create page, like this:
What I want to share with my colleagues is the results of our earlier prediction analysis—the main insight that Watson Analytics identified. That is, when the home team wins, team baseball cap sales go through the roof. To do that, I want to click on Freeform (shown in the preceding screenshot) and then on Create (in the top-right portion). For this example, we will stick to a simple single-panel, free-form dashboard.
Now we have a blank canvas to create our simple dashboard, as shown in this screenshot:
We'll perform three simple steps to assemble our dashboard.
First, let's add the Historic Sales visualization that we saved earlier by accessing the collection of artifacts. We do this by clicking on the collection icon in the top-right corner of the page, as shown in the following screenshot, and dragging the artifact onto our canvas:
Once the visualization is on the canvas, you can resize it by adjusting its borders and sliding it to the exact position you want:
Next, let's add a bit of a banner to the dashboard by clicking on the Text icon and typing our message, as follows:
Then, let's include a photo of our latest team hat direct from our supplier, by clicking on the image icon and pasting a valid web address, like this:
We can then resize and reposition our three artifacts the way we want them:
Finally, we can click on the Save icon and give our dashboard a name, as shown in this image:
Furthermore, this is too nice not to share with the team, so let's click on the share icon and select the best method for sharing our dashboard, like this:
