Bins allow you to group numerical values on a continuous metric. For example, if you are looking at the neural sentiment output for a million records, there are too many values, each between zero and one, to make sense of visually. To make it easier to visualize the data, we can create four bins in which to group the values.
- In the Widget Editor, in the Fields panel, under a Neural Sentiment Model, select the label field.
- In the Settings panel to the right, select the Pie visualization.
- From the Fields panel on the left, under Calculated Fields, drag the Number of Records field and drop it onto the Size Dimension to show how many records there are for each sentiment label value.
You can see that it is difficult to make sense of the data with so many values.
- On the Dimensions shelf, click the drop-down arrow for the Neural Sentiment Model.label field and click Create Bins.
- In the Create bins dialog that appears, click Add boundary values.
- In the Boundary#1 box, change the default 0.50 to 0.25 to set the maximum value for the first bin.
- Click Add boundary values again, and in the Boundary#2 box, set the maximum value to 0.5.
- Click Add boundary values again, and in the Boundary#3 box, set the maximum value to 0.75.
You now have four bins for the neural sentiment label values.
- Click inside the Group 1 box and add a label that is meaningful for your data, then repeat for each of the four bins.
- Click Apply to see the data arranged in the four bins.