A weekly review of news articles regarding Analytics in Natural Language
29 Apr 2019
This week we are going to have a look at Ask Data features of both Microsoft’s Power BI and Tableau.
Microsoft’s new Ask Data update called Power BI is used to support natural language queries. It’s special ‘Ask a follow-up’ allows the person querying to ask related questions. For example, If you ask ‘Top-performing products’, the system returns a set of data. If you click on the special feature and ask the next question ‘How about in Australia?’. Then it relates to the previous question and answers. The new Q&A feature also provides row-level security for all queries.
Ask Data feature is the new tableau, natural language capability. It allows the users to converse with data in published data source and can also provide visual answers. This feature is fully embedded into the Tableau server, so all we have to do is just select a data source (online too) and start asking away!
powerbi.microsoft.com The Power BI desktop has been released in November of last year and contains the latest feature to ask a related question. This eliminates the need to form a query/ ask a question from the beginning every time.
The different follow-up types of questions include Modifying the original question, Asking an Entirely New Related question, Augment the original question and putting it all together
While asking a related question, you can modify the question like ‘How many Chicago customers bought the product?’, then, ‘How many Seattle customers bought them?’. An Entirely New Related question could be ‘How many Sydney customers?’, and, ‘What did they buy the most last year?’
Augmenting the original question implies, ‘Count the customers’, and ‘Group by city’. Putting it all together refers to a series of back-to-back related questions that would be a summary of a conversation with the data source.
Q&A provides row-level security for all your data. If eastern area manager is not authorised to know about western region employees and data, then it can be implemented in Q&A.
Ask Data feature of Tableau 2019.1 beta allows the user to converse at the speed of thought thus empowering every individual in an organisation to make better data-driven decisions.
Its easy and powerful analytical capabilities allows the user to rapidly iterate over the same data set using additional follow-up questions. We can even add synonyms to the system so that Tableau can better translate our question into the precise query.
Smart Analytics feature is a pro at resolving ambiguity. Humans resolve certain ambiguity using the intonations and stresses on certain words. While the Ask Data tries to resolve ambiguity by prompting relevant questions to the user by offering helpful recommendations.
Tableau has incorporated best practices of data visualisation. Ask Data provides the best answer by analysing the context of the question and choosing the best visualisation technique. Example, ‘highest avg price by country’ automatically calls a bar graph and sorts in descending order.
This update also automatically allows for sharing of data to encourage further exploration. This can be done by saving the visualisation as a Tableau workbook share via server. Data-driven alerts too can be created.
That’s all for the week. We’d love to hear your thoughts on these articles and anything else data related! Email us at info@spectdata.com