Augmented analytics lead analytics trends for 2021

Traditional dashboards become less relevant

Analytics dashboards are a welcome addition to traditional reporting because KPIs can be monitored in timely manner, and users can interact with the data, even from a mobile device. In addition to providing role-based insights, analytical results can be narrowed further based on whom the user interacts with and business changes that are relevant to the user. However, that paradigm is changing.

"Augmented analytics is mostly being implemented as a supplement to a dashboard," said Rita Sallam, distinguished VP analyst and Gartner fellow. "I can ask the system to explain a change that's already been identified, and it'll generate a narrative that maybe explains what's going on in the data that's being visualized. I can start with a natural language question -- but I still start with a question -- and I get the answer to my question instead of point and click. But think about the possibility of it being the other way around where based on who I am, I get the five things I should focus on today. Those are more like dynamic data stories."

Platforms are evolving from being dashboard-centric to being dynamic data story-centric, and the results are becoming even more personalized.

Increased AI and machine learning services

Cloud vendors make AI and machine learning services more accessible to BI and analytics platforms and applications through open source platforms and libraries. Although content analytics, computer vision, unstructured data analysis, emotion detection and other forms of analysis are occurring in pockets, their use will become more common as more vendors consume those capabilities as services.

Sallam said "X analytics" -- Gartner's term in which X stands for anything -- will give rise to new applications and use cases that fundamentally change processes, such as taking the friction out of an insurance claim when a car accident occurs.

While analytics is already enabling many forms of optimization, Sallam said there's untapped potential as more capabilities converge in the cloud at scale and as data analytics continue to become more democratized.

Graph analytics will become more pervasive

Graph analytics have been explaining relationships for three decades, although with cloud and other advancements, expect to see improvements to natural language processing, knowledge graphs and data fabrics.

"Anything we do currently, whether it's fraud detection, anti-money laundering, drug discovery, supply chain optimization, contact tracing or customer journey analytics, it's all about finding patterns and relationships [and] being able to support wider types of data analysis," Sallam said.

Lian Jye Su, principal analyst at ABI Research, said the focus now is to apply deep learning to graph-structured data. 

IoT to play an increasingly central role

Two IoT-related analytics trends Su expects are edge-focused analytics and real-time stream processing and management.

"There are more and more data now being collected from the edge due to the proliferation of IoT," Su said. "However, many of these data [elements] come in different formats. Augmented analytics solution providers need to develop capabilities to ingest and prepare data for core analytics and [machine learning]. Developer communities are looking at Apache Flink and Kappa architecture. Public cloud vendors like AWS, Azure and Google Cloud Platform are spearheading these capabilities."

In Su's opinion, the next wave of analytics will probably focus on instant insights coming from continuous data flow, data processing automation and zero-code machine learning deployment.

More automation requires mass training

Many BI and analytics platforms now target a broader audience, which includes citizen data scientists. Augmented analytics capabilities simplify tasks such as data preparation and querying, but that doesn't mean citizen data scientists think like data scientists or analysts.

"We're not going to be training people on how to build dashboards or how to write SQL queries," Sallam said. "We're going to train consumers on how to use those insights responsibly within their context."

Citizen data scientists should understand the basics of analytical concepts and techniques including bias and its potential impacts.

To ensure trustworthy analytics, BI and analytics vendors should ensure their platforms can explain results so users understand what factors produced an analytical result. Sallam said explainability also includes proactively alerting users to potential privacy risks and bias in models.

Ambient analytics

Probably the most futuristic of 2021 analytics trends is ambient analytics taken to the extreme. To get there requires analytics capabilities virtually everywhere, embedded in all business and productivity applications and in networked devices that communicate with one another to understand the user's context, adapt to changing circumstances, personalize results and anticipate user needs or desires.

A classic example of this is a smart home. However, the concept applies to workspaces, public spaces and user experiences more generally.

"Wouldn't it be nice if I'm looking at a purchase order in a procurement application, and as I am scrolling through the different sections of the document -- customer information, detailed items, etc. -- and without asking for it an analysis dashboard​, an analysis of my relationship with the customer, an explanation​ [with] inventory availability and trends, a prediction or a recommendation just pops up?" said Boris Evelson, principal analyst at Forrester. "Or, if I get an email from you, a dashboard pops up with a complete analysis of our relationship, including the number of calls we've had and topics or trends we've discussed so all of the analysis that I need is right at my fingertips?"

Today, many applications and tools include embedded analytics designed for use within the context of the application or tool. However, these information silos can be interconnected ​via what Forrester calls data and analytics fabric.​ This is critical because to provide rich and actionable insights ambient analytics require 360-degree views of customers, products, partners and everything else.

"Enterprise-grade ambient analytics will deliver amazing results, but it'll require a significant data integration and embedding effort," Evelson said.