Augmented analytics capabilities mark the new era of BI
Augmented intelligence capabilities like automated data prep and natural language processing are now common, showing that BI has advanced to a new era of technological innovation.
After the era of self-service analytics, it's now the era of augmented analytics.
The rise of self-service analytics was driven by the idea of giving business users the capabilities to work with data without requiring the skills of a trained data scientist or data analyst.
It was about giving them tools such as dashboards and other data visualizations that enabled them to look at pre-aggregated data and make data-driven decisions on their own, in the moment, without first having to consult with data experts.
Now, however, analytics is moving beyond self-service.
Fueled by augmented intelligence capabilities and machine learning, vendors are developing tools that enable business users to do more than simply look at pre-aggregated data to inform their decisions. They're developing capabilities that enable users without expertise in data science to do some of the tasks that previously required the skills of a data scientist.
Augmented analytics capabilities are now enabling organization to develop data-driven cultures and give business users the tools to prepare their own data, develop their own data models, query their own data, build and run their own reports, and even get automated insights that lead to action.
"We're coming out of the self-service era," said Doug Henschen, principal analyst at Constellation Research, on Aug. 11 during a webinar hosted by analytics vendor Tellius. "Now, the trends are around augmented capabilities, which are bringing the ability of the computer … to the fore. This is what's shaping the market today."
We're coming out of the self-service era. Now, the trends are around augmented capabilities, which are bringing the ability of the computer … to the fore. This is what's shaping the market today.
Doug HenschenPrincipal analyst, Constellation Research
And according to Henschen, four emerging augmented intelligence capabilities in particular -- augmented data preparation, guided analysis, natural language processing and smart predictions -- are moving analytics beyond self-service and into its new era.
"Not all of these are used by everybody," Henschen said. "Some are still aimed at the traditional analysts and power users, while some are aimed at broadening the tent and getting to more business users."
Augmented data preparation
Augmented data preparation tools are capable of automating the tedious, time-consuming process of wrangling the right data for a given project, and then extracting, transforming and loading that data to make it actionable and drive decisions.
Using machine learning algorithms, they're capable of both lightening the workload for data scientists and enabling business analysts to manipulate data on their own.
"Augmented data prep is [mostly] for traditional users -- analysts and power users -- who like to get hands-on and are data-savvy and comfortable," Henschen said. "The idea is improving their productivity, helping them take on more of the data prep and data engineering tasks that would otherwise be done by IT departments."
Key features of augmented data preparation tools include automated data profiling, formatting and cleansing recommendations, data-join recommendations and data governance measures, Henschen added.
Among the analytics vendors offering augmented data preparation tools are Tableau with Tableau Prep Builder and Microsoft with Power BI's Dataflows. Meanwhile, data management vendors including Trifacta and Alteryx are automating the data preparation process.
Guided and intent-driven analysis
Guided and intent-driven analysis is augmented by analytics capabilities aimed at providing a data workflow for users who aren't particularly data-savvy.
Guided analysis tools automatically direct users as they navigate the steps of data analysis, providing a roadmap for them to follow as they explore their data with the goal of arriving at a data-driven decision.
"They're very helpful," Henschen said. "They help more ordinary business users, but also improve the productivity of more traditional users to help them do things more quickly."
Intent-driven analysis tools, meanwhile, go a step further and use machine learning to understand the habits of individual users, users within certain departments and even users across entire organizations to make recommendations.
"These are powerful features that help broaden the tent of data and analytics to more users that may not be familiar with all the nuances of exploration," Henschen said.
Natural language processing (NLP) eliminates the need to know code.
By simply typing words into a search bar or even speaking into a device, users can search and query their data and receive automatically-generated responses from their analytics tools.
The tools are able to automatically translate the natural language -- most often English but also other prominent European and Asian languages, depending on the vendor -- into SQL to run the requested search or query and then translate responses back into natural language.
"It's definitely a tent-broadener, bringing more people into data and analytics," Henschen said. "They're certainly comfortable having a Google-like experience."
NLP also includes natural language generation -- using AI and machine learning to produce narratives about data, whether data stories or short explanations of the data.
"A lot of business users aren't sure what they're looking at when they see a dashboard; they're not sure what to make of a data visualization, so natural language generation develops a paragraph describing what's important in the dashboard or report," Henschen said. "It's drawing on the metadata behind the scenes and giving a textual description."
Predictive analytics involves using the past to predict the future. Based on historical patterns, what can be expected next?
Predictive analytics, however, is complex, and has historically required data scientists to build and train models.
But now, using augmented analytics capabilities including automated machine learning, business users can use their BI platforms to look forward rather than just back at what's already happened, and do so without having to write code.
More advanced users, meanwhile, can also make use of smart predictive features and enable others within their organizations by embedding those predictions within dashboards so they're consumable by anyone who works with data as part of their workflow.
"It brings a broader base of users to predictive capabilities and predictive insights," Henschen said.
And that, ultimately, is the focus of augmented analytics. Using AI and machine learning, augmented analytics tools are designed to broaden the reach of analytics beyond trained data analysts and data scientists to give business users the power to make data-driven decisions.
