When I watched Inspector Gadget growing up, I constantly dreamed of having a book-like computer and a watch you could ask questions to and talk to people on. At an early age, I was bought into technology, gadgets and dreaming of a world where it was available to everyone. I am excited to say I now own multiple lap tops, iPads and Apple Watches and struggle to write with a pen when forced to write Christmas cards. Now that I have the platform available to use, I have replaced my Inspector Gadget dreams with the dream to more easily interact with them. This dream is fueled by recent development Tableau has recently released with Ask Data.
What is Ask Data?
Ask data allows you to analyze your data using natural language. You connect to a published Tableau Data Source, type a question and instantly get a response in Tableau. The best part is the answers come in the form of data visualizations. For a person who thinks in visualizations, this is the neatest thing since sliced bread!
But beyond how cool I think this function is, this feature is applicable for every consumer of data. It provides users with a simple and powerful way to interact with data quickly to make decisions. It really fills the demand I always get for the ability to create a table and select any dimension or metric in a much cleaner and powerful way.
To Make It Easy as Pie
As soon as we saw the beta version was available, we asked to beta test it and were quick to try it out with a few of our own healthcare data sets, OR Cases and Professional Revenue Data, once released. After doing internal testing and demoing to our entire Analytics team, we came to a few conclusions worth sharing.
To build the best Ask Data scenario, you really have to think about the data set. Often times our healthcare customers build robust applications that have 5-10 Tableau data sources. Ask Data only sits on one data source, so if you want to evaluate OR throughput times as well as OR case volumes, you need to make sure your data source blends the data sets.
You also have to consider how people think about certain concepts. For example, when we played with the Professional Revenue Data, we realized that when you use natural language, you start to use acronyms such as AR Days (Days in Accounts Receivable), ADR (Average Daily Revenue) and charges (revenue). To make sure Ask Data prompts the right scenarios and visualizes the right data, you want to alias your dimensions and metrics to support how people talk about the data.
While you may want to get straight to the most difficult question possible, such as “What is the number of cancelled cases for July 2019 with an avoidable cancel reason shown by surgeon in a bar chart,” when you start to use Ask Data, you want to start with the foundation of canceled cases and add more criteria to ensure you are getting the right data set and filtering.
Prepare for your release of Ask Data by looking at your data sources that are curated and in demand. Take a few hours to add aliases, definitions to add context for users to use the right fields and to walk users through a simple and complex scenario.
What do I dream about now?
It is always exciting to explore new technology, especially when there are such applicable use cases and value for our customers! We can see Ask Data reducing the time it takes to deliver on analytics requests and truly enable self-service on trusted, curated data sets across your organization.
I can’t wait for continued evolution of Ask Data and being able to save questions and views to my own Story or Dashboard in the future.
Want to see more?
Meet us at Tableau Conference 19 in Las Vegas! We’re attending and sponsoring the healthcare portion of the Tableau Conference this year. If you are interested, drop us a line and we can share our sponsor discount.