I would like to see the ability to group a particular set of objects based on their characteristics, aggregating them according to their similarities.
Tableau 10 recently added built-in k-means clustering, and it is really well done. I would love to see something comparable in Power BI. From the Analytics pane, you can drag Cluster into your visual to create the clusters. You have the ability to keep all continuous fields or select a subset. You can keep the automatic value of k, or you can manually select k. Once the clusters exist, you can get a description of what is included in each cluster and an analysis of variance. Automatic selection: http://i.imgur.com/rUnYF0g.png Manual feature and k selection: http://i.imgur.com/SPgUWMr.png Describe clusters: http://i.imgur.com/9fahPXF.png
