The weekend is finally here in cloudy London after what has truly been an intense week at the Data School. After spending the first two weeks on Tableau and devoting all of our attention to visual analytics and design, switching to Alteryx and delving into a brand new software proved as challenging as it was rewarding. Both pieces of software are entirely different, from the aesthetic of their user interface to their functionalities, but having spent a decent amount on each of them now it’s become clear why the DS program teaches them in conjunction. I’m going to keep this post short and go into my feelings about both of these amazing pieces of software so you get exactly why I think using them unitedly is so immensely powerful.
To illustrate what I mean, let’s picture a scenario that any data analytics consultant will have to go through at some point in his career. Say you’ve just been hired by a large retailer to help them make sense of the data they’ve collected over the years with the overall objective of identifying where they should open their new stores based on the performance of their most profitable products. As expected, their data is currently stored in .csv files hidden away in the computers of at least a dozen analysts working in different departments who use different formatting and editing rules. With only a few weeks left on your contract, you are going to have to find a way to centralize all this critical information, process it to make it machine readable and analyzable, and create useful and simple dashboards to allow future business analysts to easily explore it. Let’s take a look at why Alteryx and Tableau would make accomplishing all these tasks so much simpler than they would have been otherwise, and heck, even a lot of fun.
In a few words, Alteryx is the “back-end” of your entire data analysis process. It’s the tool you’ll be using to pull all your data from various sources, exploring it, cleaning it, and making it ready for future analysis in Tableau. With regards to our case study, Alteryx would allow you to quickly:
- Input all the different files the client has identified as key to this project and explore them very quickly using a Browse tool,
- Clean all the datasets using powerful tools like the Data Cleanser or Formula in order to remove anything that might get the way of your future analysis – punctuation, upper/lower case, blank space, etc,
- Join all your datasets together while creating additional rows or columns of data in a few clicks, whilst also easily removing any useless information. The Select, Text to Column or Crosstab tools will make all these tasks unimaginably simple.
What makes Alteryx truly impressive is the myriad of additional tools it provides to allow you to conduct some very advanced statistical and visual analytics on your data. Want to add a regression to identify which of your customer segments seem to be the most interested in a particular product? Alteryx will make that as simple as possible for you, and doesn’t even require you to really understand what is going on under the hood of any of its built-in statistical tools: you just have to run them and let the magic happen. With regards to our retail store case study, Alteryx would allow us to:
- Quickly run a series of statistical analyses on our profit data, allowing us to build a linear regression, spline model, and decision tree within a number of minutes, and providing us an easy way to compare their predictive power using the Score tool or a Lift Chart,
- Create individual location points for any of our stores for which we have longitude and latitude data, and subsequently calculate the drive time for our main customers to each store. Yes! You can do that, in just a few clicks,
- Use one of the community’s many freely-available workflows to pull in some publicly available demographic and socioeconomic data to cross-reference it with our location and sales data and automate the entire process so that nobody ever has to blend the data manually again. Alteryx will just do that for you,
- Package some key workflows that we know will be regularly used by our client in the future into Macros – you can think of them as parametrized packaged workflows with user interfaces – to save any future users hours of work.
As you can probably tell by now, Alteryx is an immensely powerful tool for both the initial steps of your analytical process, allowing you to seamlessly join different datasets and preparing them for future analysis, while also giving you access to a whole range of statistical and spatial analysis tools. After prepping our data and adding some extra fields for future analysis, you are now ready to start building some beautiful dashboard.
In your analytical process, Tableau will serve as your “front-end” software, the one you use to present the customer’s data in a way that makes it as easy to process and digest. The purpose of this software is to allow the future user – whether he be a business analyst or a store manager – to ask questions from his own data, as he will be the one best equipped to know what insights are crucially important to the business. Hence, the dashboards you will be creating must both be aesthetically pleasing and simple to get around but also embedded with all the tools one might need to effectively explore data. With regards to our example, Tableau will allow us to:
- Explore our data rapidly in order to identify key trends and outliers that we know our customer will be interested in knowing about.
- Seamlessly blend different datasets into scatterplots, maps, bar charts, or pretty much any visual tool you could imagine. These charts and plots can all be customized of course, from colors to filters, or parameters.
- Create dashboards – the key deliverable of our entire analytical process – which will serve as the main interface through which future analysts will explore the data. These dashboards can house multiple sheets, be immensely interactive, and adapted to any device (Computers, Phones, Tablets, etc.).
To the experienced user, however, Tableau can provide a whole other layer of tools that make it just the perfect business intelligence tool. With regards to our retail store example, Tableau will allow us to:
- Run some deeper analysis through the use of built-in statistical tools, as well as calculated fields, and parameters. These tools will allow you to push the limits of your data and find pivotal insights.
- Export your dashboard and data to Tableau Public, or Tableau Server, effectively allowing anyone with the right access to view the dashboard, its underlying data, to download it, edit it, make comments and suggestions. Imagine never having to send an Excel file by e-mail with 45 people CC’d, but just sharing one link to the key stakeholders?
Even though learning how to use both of these pieces of software in conjunction is as time-consuming and difficult as it sounds, one can clearly see how they augment each other’s analytical potential. From processing and preparing data in Alteryx, to viewing it and presenting it in Tableau, these complementary tools can truly revolutionize the way we think about business intelligence. Why hire 12 coders to create complicated data storage and processing centers that need to be regularly fixed when you can have one Alteryx Ace and Tableau Zen Master build you one in a matter of days that anybody can be taught to use and understand?
Thank you for tuning in again, I’ll see you all next week for more Data School thoughts and technical tips and tricks!