Have you ever seen an outlier in your data and had a feeling of dread? You may be having a great data-centric collaboration with your colleagues and you notice something out of the norm and the current dashboard might not provide the full context. The dread comes because you know you’ll have to skip lunch and work late to crawl through all of the related data, or because you're the one responsible for taking actions based on the data and you don't have time to wait for someone to come back to you with deeper analysis.
I’ve recently encountered a new Tableau feature that gives me – the analyst – a sort of superpower. That feature is Explain Data, which reminds me of an exoskeleton of sorts. It enhances the capabilities of the person using it by expediting the analysis process to provide explanations that quickly cut through the noise.
Over the course of the last few months, I’ve been playing with the beta version of Tableau 2019.3, which includes Explain Data. In that time, I’ve found that Explain Data helps me take care of the busy work by interrogating my data sources and all the related dimensions so that I can quickly apply my subject matter expertise and critical thinking to identify any actions I need to take. I always think it’s easier to understand a new feature when you see it in action. Let’s dig into a few use cases I’ve discovered in my role as a BI consultant.
Use case 1: Uncovering the rise in academic enrolment
Throughout the United States, there has been a lack of clarity for decades regarding the actual status and condition of arts education courses in K-12 public schools (from nursery to secondary school). The Arts Education Data Project is there to acquire and integrate data from disparate data sources, so that schools and policy-makers have a means of benchmarking and to optimise existing programmes. In support of this goal, it is important to understand the details behind longitudinal changes. For example, if participation in a particular arts discipline is going up, what possible factors may have contributed to that increase? Is it something that should be further promoted? How can we learn from it? With limited resources to research answers to these questions, it is important that we are able to quickly and efficiently understand all of the longitudinal data and related dimensionality to gain insights into the most important driving factors. Enter Explain Data.
The following dashboard illustrates longitudinal data for the state of California. At the bottom of this dashboard, we see that enrolment on the “Other Arts Course” discipline is on the rise but it is not clear what is driving this trend.
By using Explain Data on the most recent mark in this sheet, we can see the courses that compose this discipline and their enrolment trends.
In this instance, the visualisation generated by Explain Data lacks some clarity, but the beauty of it is that it is seamlessly integrated with Tableau, so we can pop out the visualisation to explore it further from its own standalone sheet. From here, we can add a sort and now we have useful insight that the course “Intermediate Graphic Design” is the single largest contributor to the increased enrolment.
Use case 2: Understanding outliers in speech sentiment data
Historically, speech analytics has been thought of as being confined to the contact centre and limited to superficial metrics such as number of calls and length of calls. However, with modern transcription technology and open architecture solutions, the true voice of the customer can be understood and integrated with other enterprise data. This is made possible by Tableau Technology Partner, VoiceBase. When the audio file from the contact centre recording is processed, an enriched transcription is produced that includes data beyond the words that are spoken, to include things such as categorisation of what was discussed and the sentiment of each person. Was the customer angry when they called, and ended the conversation on a positive note? The list goes on. This opens doors for potential applications across the enterprise. Although this presents exciting opportunities to leverage data in new and innovative ways, it also presents more and new data to see and understand.
The dashboard below provides insights into why people are calling a contact centre. This information provides organic feedback that is applicable on a number of fronts. It can be as basic as managing contact centre staffing to understanding how more proactive solutions can be provided to better serve customers. The data source for the following visualisations is “Thrive Air”, a fictitious airline akin to Tableau’s Superstore.
Digging into sentiment, we quickly see that lost luggage is problematic. What might be causing this low sentiment on behalf of both the caller and the agent in the call centre?
Running Explain Data on the mark of this outlier interrogates all of the other related data available in the corresponding data source, including dimensions that are not included in this dashboard. From here, we can see that one of our call centre managers, Tom Taylor, has an opportunity to better train his staff on how to resolve lost luggage calls.
Enhance your analytic capabilities with Explain Data
As you can see, Explain Data enhances your analytics process by assisting you in your analytics journey.
Explain Data could change workflows in a similar way to how Tableau Desktop originally did. Tableau Desktop enabled access and exploration of data to more people without waiting for the classic report factory. With Tableau Desktop, the user might not have to write code. Likewise, Explain Data will enable new segments of users to dig into their data. With Explain Data, the user might be able to answer “why” questions about their data without even having to drag and drop a pill.