A Pratical Look at Qlik & Tableau
by Marina James, on Oct 16, 2018 10:43:44 AM
Data visualization is a very mature space to say the least. Most companies have multiple data viz options. These tools and platforms can be amazing problem solvers turning analysts into heroes and transforming how the business measures success.
There are some pretty loyal camps in the data viz space, but it’s important to remember the tool should be chosen based on the need. Not the other way around.
Also, how do you know which tool to use if you have never applied it in a real-world scenario. The only way to understand how to best leverage your panoply of tools is practical experience.
As our customers trusted advisor, we must have deep tool expertise, so we can understand what problem they are trying to solve. Only then can we provide the right approach.
Recently, I was tasked with migrating an application from Qlik to Tableau, keeping as many of the functionalities present in the Qlik version while also adding in some interesting Tableau novelties to help the end user get the best possible post-migration experience. This blog post goes through my experience building a Qlik application in Tableau. I will highlight the major differences in each platform as they pertain to this use case. The data included in the mock-up is randomized data.
My intent is not to support one tool over the other; rather, it is to provide an overview of my experiences and opinions using both tools to give you a better idea of what each has to offer. Also, this article does not cover the entire platform of either tool. This post just covers visualization, and other core features such as security, embedded analytics, etc. should be consider by any company doing an evaluation. The right tool for your organization could be Qlik or Tableau… or MS Excel… or any other tool. This world can get confusing, so if you have questions, please reach out to email@example.com to get more information. If you have any comments, please feel free to leave them below.
View our Applications
Want to see this application in Qlik Sense? View our Qlik Sense application here.
Want to see this application in Tableau? View our application on Tableau Public here.
Data Viz Philosophy
Migrating applications between tools is not as straightforward a task as one may think, though it is not impossible. Qlik and Tableau are both focused on different benefits for the user: Qlik prides itself on its associative data engine, allowing users to glean information by seeing the connections between different data points, while Tableau promotes its “question everything” mentality, encouraging users to foster new questions from what they see on the screen and easily create the visualizations that answer those questions. In short, Qlik focuses on data relationships, and Tableau focuses on ease-of-use.
Commissioning a platform change from Qlik to Tableau (or vice versa) is a major endeavor and should only be undertaken after serious thought and planning. Each platform offers users many intriguing features that are not present in the other and being aware of these limitations is a crucial first step to ensuring as few hiccups as possible.
“Put the power of data into the hands of everyday people, allowing a broad population of business users to engage with their data, ask questions, solve problems, and create value”
-The Tableau Mission Statement
Tableau focuses on “empowering” the average user, giving them the resources to investigate their own data, come up with individual insights, and facilitate the development of new questions to gain new insight. By empowering the average user, Tableau strives to reduce the burden on IT and make the average user more powerful. This mission statement is a direct result of Tableau’s overall philosophy: questions are good. If software can help a user develop more insightful questions to gain as much information as possible, and then answer those questions, everyone benefits.
“The problem with most BI vendors is that they rely on query-based analysis that restrict people to linear exploration within a portion view of their data. Qlik’s associative engine lets you combine any number of data sources so you can freely explore across all your data…”
-The Qlik Associative Difference
Qlik, on the other hand, focuses on providing that special look into data, as the data relates to other data points. Qlik aims to give the user the ability to see how dimensions and measures are connected to other dimensions and measures, allowing them to glean as much insight as possible from what they see on the screen. By exploiting the inherent association within the data model, Qlik moves to close the blind spots that may come up when the whole picture cannot directly be seen.
Migrating the application across platforms was an interesting exercise. Being able to quite literally compare the differences side-by-side allowed me to better understand the platforms and the effects on the end user. The top things I learned include:
- Importing data into Tableau requires a different mindset than when importing data into Qlik—all transformations need to be accomplished prior to importing data, whereas Qlik encourages transformations in the program itself
- Tableau is better designed for ad hoc queries, while Qlik offers more functionality for static reporting
- Tableau visualizations differ in overall quality than Qlik visualizations, generally for the better
- Tableau was generally easier to learn, but becoming an expert can prove challenging
- The Qlik scripting engine and custom integration (API) capabilities are a little ahead that Tableau. However, Tableau’s release of its data prep program (Tableau Prep) and Custom Objects in 18.2 close the gap significantly
I did not get into the administration, security, or application scheduling aspects of Tableau for this exercise. Both Tableau and Qlik have different pro and cons in this area for enterprises.
Bringing in the Data
Qlik was built to give a developer the tools required to transform data into a form that worked best with how analyses were to be conducted. As such, Qlik created a product with two parts: a back end utilizing a SQL-like syntax made for SQL-like transformations, and a front end that housed many exciting visualizations and included the capacity to extend functionality with extensions.
Tableau, though, took a different approach. Instead of focusing on getting the data in the right form, Tableau initially relied on the original data source to output data in the correct form. Tableau was just a front-end visualization machine. As business needs have changed, so too have Tableau’s offerings. Tableau now offers a new product (at the time of this writing) dubbed Tableau Prep, which aims to help developers clean their data and prepare it for use in Tableau’s front end. This product is still in its infancy, however, and a number of features can still be developed further. It is essentially a lighter version of Alteryx, but native to Tableau. It’s a BIG deal, and potentially a game changer for Tableau.
Because Tableau focused more on the visualization portion of Business Intelligence, the method with which one brings data in differs significantly than with Qlik. With Qlik, a developer can bring in a number of tables, perform transformations, and run the script. Qlik’s associative engine is able to associate fields from different tables with one another by matching like field headers. By doing so, Qlik can create a robust data model from several different tables in the program itself. Tableau, on the other hand, prefers a single fact table source to work at its most efficient level. It does have the ability to perform joins and unions (concatenations), but in my experience, I have found that utilizing these options significantly affects rendering times for the visualizations on the front end.
In Qlik, the modeling is housed in Qlik’s proprietary format. Tableau’s reduced ETL capabilities forces users to make changes outside of the Tableau interface in the data layer. Overall, data management at the database, data warehouse, or data lake levels is a better long-term enterprise strategy as it becomes easier to source from other applications. Qlik’s scripting allows users to move quickly, but over time, ETL logic tends to prefer logic being moved to data delivery systems. It will be interesting to see how Tableau Prep handles this in the coming years.
Interpreting the Data
While Qlik and Tableau both advertise catering to the same audience (business professionals who want to get more information from their data), their different approaches create a very different experience for the user. From my experience, I see Tableau more for ad hoc queries, empowering the user to answer their questions on their own. Qlik, though, is a different story. QlikView is geared towards generating more static reports, with the onus on the developer to utilize the advanced functionalities to create an informative product. Qlik Sense was designed to be a hybrid, giving a developer the tools necessary for advanced visualizations while giving the general user the tools necessary for ad hoc queries. In practice, Qlik Sense swings a little more towards a face-lifted QlikView.
Visualizing the Data
Tableau’s visualizations are seriously some of the best out-of-the-box representations I have seen. They are mature enough to be used in a professional setting, but whimsical enough to be interesting. To me, the crisp, clean charts and graphs are pleasing to the eye and convey information in a non-obtrusive way. Mimicking these visualizations in Qlik (either QlikView or Qlik Sense) may take some concerted effort on the part of the developer, as a number of native visualizations in Tableau require extensions in Qlik. On a spectrum between UI design in QlikView and Qlik Sense, I would put Tableau somewhere close to the middle, leaning towards Qlik Sense. QlikView’s visualizations are very mature and quite plain. They convey the necessary information without much flair. Qlik Sense, on the other hand, takes a more modern approach with the visualizations.
Tableau also offers a number of benefits over Qlik in the realm of visualizing data. Tableau’s engine is able to associate location dimensional attributes with an actual location, effectively removing the extra step of importing location data to tag the dimension. This removes the burden on the user, allowing them to simply drag and drop in a country, state, province, zip code, etc., and move forward with analysis. Additionally, Tableau offers users a “Show Me” menu to help users decide which visualization could represent the data they have. All users have to do is drop in the necessary dimensions and measures into the canvas, and Tableau will offer suggestions for potential visualizations that could represent that data. It is a very great starting place when you have your data, but do not know how to convey that information.
Learning from the Data
With this client, Qlik was the main visualization and BI tool for a number of years. As such, users had grown accustomed to what Qlik had to offer, namely looking at data with the “green-white-gray” association model. Qlik’s Associative Engine, the driver of the “green-white-gray,” is a clear advantage over Tableau’s “non-connected” view into the data. Qlik users felt at home being able to click on certain data points and seeing everything associated with that data point. Moving to Tableau forced these users to learn a new tool that was fundamentally different. Gone were the automatic filtering capabilities that spanned the entire workbook and the default ability to select a dimension on a graph and have those selections propagate throughout the entire workbook. These functionalities are available in Tableau but require some additional development work to be fully realized. Additionally, Qlik does offer developers more advanced analysis through their set analysis feature. Tableau has been ramping up its analysis offerings, with level-of-detail calculations being introduced in Tableau 10. Overall, starting with Tableau from the beginning provides a much easier learning experience than starting with Qlik, growing accustomed to its features, and then having to learn Tableau.
In my experience, Tableau has created a platform that encourages asking questions to get as much information as possible. Tableau’s interface may require some learning, but a little training can help a user generate very insightful visualizations.
Many organizations support both, and platform switching should only be considered if the benefits far outweigh the costs.
Krishan Patel is a Solutions Consultant at Axis Group. Krishan holds a bachelor’s degree from the Georgia Institute of Technology in Chemical Engineering, and he is working towards a masters degree in Analytics from the same school. Krishan is one of the first consultants to adopt Tableau as part of Axis Group’s new platform-agnostic approach to data literacy, exploration, and analysis. He strives to help his clients “see the meaning behind the numbers,” helping them understand the data they have and how they can best work with it to gain insight and make informed decisions.