With the overture of Natural Language Processing in Business Analytics, Ask Data feature which is a part of Smart Analytics Capabilities offered by Tableau in Tableau Server and Tableau Online (released from version 2019.1) allows users to interact with the data and gain insights. This unveils countless capabilities in business analysis and leverages the self-service feature of Tableau. It lets the user ask questions to the published data for which the responses are provided in the form of visualizations which helps in arriving at data-driven decisions by providing the ability to explore the data. This blog from the series of Smart Analytics in Tableau gives a swift preview of how to prepare and Ask Data in Tableau.
Preparing Data for Ask Data
Dealing with the data, for the Ask Data feature to give accurate output we require the data being used for it to be orchestrated to support ideal analytical conversation. Ask Data resolves the user’s intent by aggregating the measures based on the data type in the source and the keywords used in the question to specific attributes. Resourceful output can be provided by the inference algorithm when the attributes are of expected data types. Having a minimal data source with only the fields that the user might use for querying will improve the initialization and parsing performance, thereby ensuring the user’s good performance. The measures must be set up with appropriate formats (e.g., currency format) if applicable. Logical hierarchies can be established to allow the user to drill up and down with the visualization created by Ask Data. Naming the fields with unique and meaningful names can avoid confusions that may affect the visual or insight created and avoid ambiguity with data. Since Ask Data cannot create measures on the fly, calculations must be created prior to using the data source to allow the user to question about those fields. Synonyms can be added for the fields that can help in identifying the appropriate field to arrive at the output.
How to ask your data?
No additional setup is required to dive into the smart analytics feature of Ask Data. You can just select a data source and start questioning straight away with simple statements and questions. The semantic models by Tableau enrich the data for more competent interactions. Ask Data can help you out with asking questions by suggesting questions based on the chosen data source. It understands ambiguous statements and resolves them by giving recommendations. The resultant visualizations from Ask Data are compliant with best practices, which are limited to a fixed number of visualization types. You can choose to save the result to a Tableau Workbook and then modify the visualization to all possible types.
Follow the steps below to ask your data and gain useful insights:
Select any published data sources from the list available and click on the ‘Ask Data’ tab available on the left side. It lists all the fields available in the tab on the left and the area where you can ask your data.
Ask Data also suggests a list of questions based on the available data from which you can select. Tableau answers your questions by displaying a visualization.
The type of visualization can be changed from the list of options available as shown below:
You can make use of the visualizations by saving it as a separate workbook for later use where you can perform analysis and modify the visualizations with more options available for workbooks. Ask Data allows easy questioning of data making it a great leap in the self-service analytics offering. Check out the blog on Smart Analytics in Tableau to know about all smart analytics features provided and reach out to us to analyze your data and gain productive insights.
To learn more about Visual BI’s Tableau Consulting & End User Training Programs, contact us here.