Sergey Nivens - Fotolia
Analytics and BI platforms have entered a new era, thanks to AI and machine learning. Augmented analytics tools help data analysts, business analysts and end users surface better quality insights faster and easier.
"You could make the argument that there's no better time to have augmented analytics," said Kjell Carlsson, principal analyst at Forrester Research. "It will be all the greater in these uncertain times because you can understand how your customers are changing or have advanced warning when your supply chain is getting out of whack, which it naturally will be."
Augmented analytics tools had a greater impact on the 2020 Gartner Magic Quadrant for Analytics and Business Intelligence Platforms this year than in 2019. For one thing, capabilities are more developed now.
Two of main features are natural language querying and natural language generation capabilities. Natural language querying allows users to surface insights without needing to understand a query language, such as SQL. Natural language generation allows platforms to respond to queries in plain language and "narrate" visuals so users can better understand what the data is saying.
Increasingly, BI and analytics are becoming more conversational, although querying capabilities tend to be text-based rather than voice-activated. Users may still have to adjust the form of their queries to the limitations of the platform, though capabilities will continue to become more sophisticated over time.
"In Tableau, instead of dragging and dropping the various variables, you could query it as you would a search engine like Google, and it would not just pull that data back, but then it would automatically generate visualizations that were applicable for understanding the data," Carlsson said. "Those capabilities are not about analytics or AI, per se -- it's more cognitive search or more rules-based automation techniques than we've had before."
Another augmented analytics feature is automated insights, which can take different forms. Qlik supplements queries with suggestions and surface insights that the user may not have considered. Sisense can surface insights based on a user's role, the industry the user worked in, and what similar users have done.
"We're seeing a shift in the user experience from one that was dashboard-centric to one that is far more dynamic and revolves around dynamically generated data stories," said Rita Sallam, analyst and fellow at Gartner. "So, instead of going to a dashboard every day or every hour to see what's going on in the aspect of the business I care about, I might get these dynamic data stories that get to the punchline."
Another feature of some of the platforms have implemented is automated data preparation.
"Augmented analytics is essentially leveraging machine learning and AI techniques to automate or augment the tasks involved with preparing data -- profiling it, cleaning it, joining it -- to automate the tasks around generating insights and explaining insights; and in doing so, it relieves the business person who is maybe an analyst or a consumer from having to manually explore data," Sallam said.
However, Sallam warns that augmented analytics may give organizations a false sense of security because they're easier to use than previous versions of the platforms.
"There's sort of a paradox where we think that the more we automate, the less we need to train people, but it's actually the opposite," Sallam said. "The more we automate in terms of insight generation, the more we need to teach consumers how to leverage that power within their context. And so, our view is that its full potential requires commensurate investment in data and analytics communities and user enablement and literacy to ensure that those capabilities can be leveraged to their full potential."
One of the criteria Gartner used in its 2020 Magic Quadrant is making sure users can understand how an insight was generated, whether there's potential bias in the insight or potential privacy violation risks with the underlying data.
Power BI targets a billion users
Power BI has an unfair competitive advantage because it has the strength of Microsoft behind it. Specifically, Azure, SQL Server and Microsoft Office may collectively help the BI group realize its goal of one billion Power BI users.
The BI group also includes AI advancements that 3,000 to 4,000 individuals on the AI team are making in object recognition, speech recognition, machine translation and machine comprehension. Arun Ulagaratchagan, corporate vice president of Power BI, said his team takes those innovations and create business user experiences that take advantage of AI.
"Power BI [has been] a cloud service from day one," Ulagaratchagan said. "It can connect to any data source on premises or in the cloud, and you can consume analytics on any device, as embedded, or on HoloLens."
Power BI currently supports 43 languages and has 40 primary and backup data centers around the world. The platform also has just about every type of certification one could imagine. Some of the other benefits, according to Ulagaratchagan, include:
- user interface affinity with PowerPoint and Excel -- sharing the same ribbon as Microsoft Office, so end users don't have to learn anything new;
- high mobile app ratings;
- native integration with Microsoft Teams for secure collaboration -- Microsoft Information Protection allowing users to label a report confidential or export encrypted data into PowerPoint, Excel or PDF as per Active Directory credentials;
- Power Apps low-code application development and Power Automate low-code business process automation;
- more than 150 connectors;
- built into Azure Synapse Analytics, the next generation of Azure SQL data warehouse, with the ability to build reports within Synapse Studio, so customers can build out components of the large BI systems in one place; and
- competitive pricing: The authoring tool is free, Power BI Pro is $10 per user, per month, and Power BI Premium is $4 to $5.
Qlik suggests what you haven't considered
Josh Good, vice president of product marketing and data analytics at Qlik, said the company has been working on augmented analytics tools for three or four years. One of the things the company noticed was that machine-generated insights often lack context and human intuitions.
"With Qlik bringing augmented [analytics] into things, we very much look at this as a way of engaging people more in analytics, surfacing insights they may not easily find [and] upleveling their analytical capabilities," Good said. "We talk about augmented analytics being a vector for data literacy."
Qlik's Cognitive Engine works in tandem with its Associative Engine, which understands the data. If a user asks about last quarter's sales, they won't get a simple answer. It will say, "Here's your sales trend, and here's what your sales trends look like by region."
Qlik shows data from multiple perspectives, helping users think more critically about data analysis. If the platform shows a bar chart, but the user would prefer a pie chart, the platform won't just generate one. It will inform the user why the bar chart is recommended over the pie chart.
One benefit of Qlik's augmented analytics tools is that users aren't limited to the scope of their queries. According to Good, other benefits of Qlik are as follows:
- Qlik doesn't care where your data is, the company doesn't keep your data and it does not claim ownership of your data in its system.
- It offers user-friendly web-based and desktop tools.
- The data stays secure on the infrastructure of your choice.
- The Associative Engine maintains the context of the query, so users get the answer to a query and other relevant insights, including multiple data visualizations.
- It can integrate with third-party speech recognition tools (Amazon, Google, Apple).
- Qlik supports third-party natural language generation; however, Qlik is also developing its own.
Tableau is evolving
Tableau, like Qlik, has been a major force in the BI and analytics space for many years. Last year Salesforce acquired Tableau, but the companies have not revealed what impact the acquisition will have on individual products. Historically, Tableau was associated with awesome data visualizations, but since that capability has been commoditized, Tableau continuously refines user experiences.
"Tableau has a singular focus to help people see and understand data," said Charles Schaefer, senior manager of competitive intelligence at Tableau. "This has been our mission since the company was founded, but it has never been more important than right now."
Some benefits of Tableau, according to Schaefer, include:
- the Ask Data natural language querying capability allowing users to type a question and receive an interactive visualization;
- Explain Data taking customers from "what" to "why" by automatically uncovering and explaining the value of specific data points and automatically explaining the most relevant factors driving a given data point;
- voice interaction on the road map, which integrates leading voice services such as Salesforce's Einstein Voice Assistant;
- extensions to AutoML solutions, including DataRobot, Aible and Ople; and
- augmented capabilities baked into the entire process -- from data preparation and management to recommendations that guide users to the right content.
Sisense: For sophisticated builders
Unlike Power BI, Qlik and Tableau, Sisense is targeted at more sophisticated users than business analysts or business professionals. It allows "builders" -- BI teams, product teams, cloud data teams -- to create custom data experiences for their users.
According to Harry Glaser, chief marketing officer at Sisense, the company has a well-staffed AI lab led by well-known research so its platform can augment every part of the product with AI.
Some of Sisense's benefits, according to Glaser, include:
- auto predictions on graphs so historical trends are supplemented with a prediction that includes error bars;
- auto deduplication in modeling -- for example, if one column says "SF" and another says "San Francisco," the platform will automatically resolve the issue;
- natural language querying, giving users natural language answers instead of interacting in a point-and-click fashion;
- a built-in knowledge graph Sisense has developed anonymously tracking customer usage so it can anticipate insights based on a user's role, industry and type of data peers typically look at;
- integration with Amazon Alexa;
- partnership with Narrative Science for natural language generation -- although Sisense is building its own capability; and
- automated data preparation.