Clustering and other Features of Tableau 10.0

Step 1:- Open Tableau 10.0 Beta and connect to “Real Estate” and “Real Estate – Supplement” data sources .

Here we use tableau 10.0’s one of the new features i.e. Cross Database Join. Tableau automatically finds the connection between two data sources and uses inner join.

1

 

Step 2:- Now move to the sheet1, Add Area and Price to the columns and rows respectively. Then insert address to the Marks(Details).

Now we move to the Analytics tab and there we can find a new feature CLUSTER under Model section. Drag the Cluster from model to the screen. We can see that Tableau 10.0 automatically creates appropriate clusters for the given data. However we can manage the number of clusters we want and also use different measure’s to get clusters according to our requirement.

4

 

 

6-ii

 

In the above picture you can see that we have selected number of Bathrooms and Bedrooms gives different clustering results.

Step 3:- Now drag the cluster from the Marks section to the dimension so that we can use it in other sheets. Again drag it from dimension to the Marks section and replace the first cluster. Now we can rename the clusters by selecting edit alias option.

7

 

 

Step 4:- Now create another sheet and name it sheet2. Add longitude and latitude to columns and rows respectively. Now drag “Price” from the measure to the Size in Marks section and the cluster that we have created from the dimension to the Detail in Marks section.

8

 

 

Step 5:- Tableau 10.0 has also added one new feature in terms of selection in the maps i.e. radial selection. Now we can easily find our point of interest from the desired point within specific radius.

9

 

 

Step 6:- Now go to the Analysis option and select Highlighters option and check mark desired options. Now Highlighters are added to the screen and here we can find our required points easily among all other points and compare them easily.

10

 

Leave A Reply

Your email address will not be published. Required fields are marked *