Having recently spent an evening with Tableau, I decided to share my notes. Many of the features they highlighted I did not know about. Here’s a short list of cool things to do in Tableau. All of the visuals link to the actual Tableau example used.
On your visualization, you can select multiple items and create a ‘group’ of these to differentiate into larger sets. Sounds simple enough of a feature but it allows some pretty nice things to be done AFTER the data has been set, schema and all. It allows us to maintain additional attributes/additional dimensions if you will at the Tableau layer instead of enriching the data itself.
We can go from this
1. Draw a box or do multiple select to define the items you want to group together. You would be doing this for every group you want to define.
2. Having all the group members elected, right-click, select Group
3. To create a label for a group you have already created, just right-click the group and Edit Alias
This is very neat. Instead of having to merge different data sources at the data layer, we where shown by example how to add and additional datasource to an existing visualization. I assume Tableau does a merged table of data or perhaps some join between these. Things happened too fast so I cannot tell what actually happens. At least for sensibly related data, this can potentially save a ton of time. Very neat!
Another neat feature is the ability to add annotations to the visualizations. Simply right click, Annotate an element on visual. Multiple elements can be annotated and there are various options depending on what you want to annotate.
Easily forgotten, Tableau allows the creation of derived metrics in a few different ways. This possibly saves an aggregate table or two, and definitely a lot of planning.
This one feature I thought would be goofy to even try. I think I was mistaken as it does provide striking visuals without busying display too much. I guess the ability to combine more than one visual display can be a very effective way of expressing facts.
Here I combine a ‘heat-map’ and a bar-chart. The end result does indeed convey both prices per London Borough AND the aggregated price per Area (defined as a group!!) effectively.
Embedding Visualizations In Content
The ability to embed Tableau visualization in your own content (Tableau server or Public hosts the interactive visuals for you). In my case, I cannot show this because Free WordPress won’t allow it. Just head over to Tableau’s Gallery for plenty of examples. Layouts look very clean.
Binding Controls (for all visuals in dashboard)
This is the ability , in a dashboard, to click on one visual and selecting that group in all the visuals in the dashboard. Think about filtering the set by clicking on one element on one sheet and all the other sheets on dashboard filter out as well.
To do this in a dashboard, select an element from a sheet to filter on, then right click and choose Use as Filter from the menu.
From here on, selecting a group, in this case, will filter out set both in the sheet you clicked and in the other sheets visible in the dashboard.
Eliminating Cubes (OLAP)
Tableau also brought a local customer that went thru multiple use cases they use Tableau for. One of the most insightful points I gathered was that they can eliminate maintaining cubes for their users. This itself eliminates a ton of grunt work. Since the first step of everything in Tableau, at least in the Public version, is to import the datasource, cubes are negated. Tableau ‘engine’ replaces this requirements instead. A combination of database and Tableau Server planning would have to be thought of; there is no free lunch but this is a very neat change in workflow.
Working On The Browser
At one point, the Tableau guy (hi Shawn!) was browsing a Tableau Dashboard and proceeded to edit the mechanics of it. I immediately thought Tableau on Chromebook! but could never get this to work at home. Maybe this is a feature of the hosted product and not the public version.
There are many features non traditional products like Tableau bring to the customer. Many of which make the daily life of an analyst better. I think it empowers a lot of users that may shy away from delving too deep into using visuals to convey information. If anything, it provides a fresh perspective on the analytics landscape companies are accustomed to. This is worth considering. Thanks for reading!