Analytics
News, events, thought leadership and more.
Last week we held a webinar with our customer amaysim of Australia, where they presented their use of Alteryx, Tableau, and Amazon Redshift. Julian Dell, IT Director, and Adrian Loong, BI Manager, spoke about how they chose this stack, how they have driven executive support, and the cultural impact on their organization of becoming an analytics driven organization. You can hear the recording of the webinar or see the slides on Slideshare.
As we mentioned in the webinar, we wanted to answer the questions we received in the webinar in a blog post. So below are some of the key questions and answers from the webinar.
Where is the "blended" data stored? Redshift?
- Brian Dirking: The blended data can go back to Redshift, or go directly to Tableau as a TDE file. The beauty of using Alteryx, Redshift, and Tableau is that you can do blending and advanced analytics in Alteryx and put the results in Redshift, and let Tableau pull the data directly from Redshift, or have Alteryx pull data from Redshift and blend and apply advanced analytics to the data, and then provide the results to Tableau in a high performance TDE file.
What does #3 mean? Does that go beyond writing custom SQL statements? (In the amaysim slides there was an image of the architecture that included text that "Alteryx can apply and validate more complex business rules before visualizing outputs in Tableau.")
- Brian Dirking: Using Alteryx, you can build the blending without having to write custom SQL statements. Alteryx provides an easy to use interface to blend the data and apply business rules. You can apply those business rules in a couple of ways. You can have the analyst who is creating the workflow create and apply those rules. But furthermore, once you have applied them you can turn that set of macros into something that can be called and used over and over again in every workflow you create. You can even black box it so it can be locked and protected from user access.
Why wouldn't you use redshift for visualization? [It] seems it has the scalability to support large dataset visualization.
- Dustin Smith: Many people do exactly that by connecting Tableau directly to Redshift. Still others will use Alteryx to blend and enrich data and then put that data back into Redshift and then connect Tableau. It's all about options.
In getting folks to use the tools themselves, what did you do to make the different data sources understandable to the users who are not familiar with the data?
- Dustin Smith: Many people use Tableau Server's "Data Server" capability to create a data portal for data users. Friendly naming, notes, pre-built hierarchies, etc. http://www.tableau.com/learn/tutorials/on-demand/data-server
- Brian Dirking: Furthermore, you can use Alteryx to normalize those data sets so they become consistent. You can retitle field names, reorder content, combine it – all through a drag and drop interface.
What is the process time in Alteryx? Let's say you have 10mm rows and you want to apply several 'filters'?
- Brian Dirking: There are lots of variables there but I would say that customers are constantly surprised at the speed of Alteryx for processing. It is used in high volume data environments by Experian and Rentrak. You can see the customer at Rentrak speak about how when he moved from SAS to Alteryx he was amazed at how much faster Alteryx was to process the same jobs.
Is there a place for intro training on data blending in Alterix? Does Alterix have month-by-month cloud-based service plan?
- Brian Dirking: Alteryx training can be accessed at http://www.alteryx.com/product-training. There is a fair amount of training online. Alteryx Designer is licensed annually, and Alteryx also offers the Alteryx Gallery for cloud-based access to analytics applications.
Did you use any consulting partner to help you in this process? or was it all done by internal resources?
- Brian Dirking: Julian mentioned their partner MIP, who introduced them to Alteryx. They continue to partner with MIP through this whole process. (At the end of the webinar Julian also mentioned that they are going to work with MIP as they go deeper into predictive analytics.)
What percentage of your team time is spent on Alteryx vs Tableau?
- Dustin Smith: I can't answer specifically for Adrian and JD, but it varies greatly by person and role. For folks that use both products it's largely 50/50. Blending/enriching leads to more visual discovery which leads to more blending & enriching.
Are you able to do query concurrencies via the tableau server?
- Dustin Smith: Is the question "Can Tableau Server handle concurrent queries while the dashboards are connected live to Redshift?" If so, the answer is yes.
When using the analytic features of Alteryx and visualizing in Tableau, does Tableau use the Alteryx engine to perform the calculations in real time or do you build the data frame in Alteryx and then visualize on the static data frame?
- Brian Dirking: You can do this in a number of ways. Alteryx can provide the data and Tableau can calculate, or you can calculate in Alteryx and deliver the results to Tableau using a TDE file. It depends upon your use case which one you would choose. Lots of options!
