 We love our data, but most of the time our data is big with many variables, and making sense of it is a difficult task. Well, not so difficult in orange. Let us load the iris data that we have used in some of our previous videos and connect it to Scatterplot. Scatterplots are great. They can show us the relationship between two variables. But our data rarely has only two variables. What an iris data has for? How could I see the relationship between more than two features? With freeWiz. FreeWiz is a multi-variable projection method for uncovering feature interactions in a data with class variable. If we pass iris data to freeWiz, the widget would initially place the variables on an invisible circle, making them equally important. Just optimize, and freeWiz will rearrange the projection and find the one that best separates points of different classes. By doing this, optimization exposes the most relevant features of the data. Variables associated with longer axes are more important for a specific class value. And those variables that lie closer together are more correlated too. FreeSitosa flowers, which are marked with blue circles, have high values of sepal width. And large petals are distinctive for iris virginica. On the other hand, sepal length does not play any role. Let us see this on a different example. The zoo dataset contains a hundred animals classified into seven groups, including mammals, fish, bird, and so on. We will use FreeWiz projection to investigate relationship between all 15 variables and see whether there's some structure in the data. In the file widget, simply change iris to zoo with browse documentation datasets. In FreeWiz, we instantly see the unoptimized projection. Let's use optimize again to reveal the relations between features and animal classes. You can even use show class density to observe whether the data projection in FreeWiz really managed to distinguish between different class values. It looks like it does. Having hair and giving milk is a distinguishing property of mammals, while being aquatic is a property of fish. On the other side of the projection are those animals that lay eggs and have feathers. This means mammals do not have any of those two properties, since they are placed opposite of the two features. Finally, at the center of the plot, we see features that are not very informative, such as animal domesticity. We can exclude these features from the graph by increasing the blackout radius. Visualizations make data interpretation so much easier. In this video, we learned how to uncover interesting relations between classes and features in a class label dataset and how to interpret the FreeWiz projection.