Augmented Analytics: Unlocking the insights from your data

This Industry Viewpoint was authored by Rohit Maheshwari, Head of Strategy and Product at Subex

Big Data is a big business and it’s getting bigger with each passing year. In fact, IDC predicts that data generated by connected Internet of Things (IoT) devices will grow from 13.6 zettabytes (ZB) in 2019 to 79.4 ZB by 2025. Although Big Data plays a key role in an organization’s ability to make effective business decisions, most companies struggle with being able to accurately interpret the insights held within that data. There are two reasons for this: the sheer volume of available data, and that most organizations still rely on processes that are manual and prone to bias across the entire data value chain.

The challenges of unlocking the value from data were highlighted by Forrester Research estimates showing that less than 0.5% of all data is analyzed and used, and only 12% of enterprise data is taken into consideration when making business decisions. Even more concerning is that Forrester believes that a typical Fortune 1,000 company increasing their data usage for decision making by just 10% could yield more than USD 65 million in additional net income.

Augmented analytics can hold the key to addressing the challenges organizations face in uncovering the insights and maximizing the benefits of Big Data.

Not a new concept, augmented analytics was defined by Gartner in 2017 as the use of enabling technologies such as machine learning (ML) and artificial intelligence (AI) to assist with data preparation, insight generation and explanation to augment how data is managed in analytics and business intelligence (BI) platforms. Gartner goes on to explain that augmented analytics also helps both expert and citizen data scientists by automating many aspects of data science, ML, and AI model development, management and deployment.

Comparing the current approach to augmented analytics

With the explosion of data, many organizations are finding themselves drowning in a sea of Big Data. The current approach to garnering insights includes managing and preparing the data for analysis, building AI and ML models, interpreting the results, and creating actionable insights. This approach leaves business users to find their own patterns, and data scientists to build and manage their own models. The manual effort of traditional analytics, more often than not, results in users and data scientists left to examine their own hypothesis. This can result in key findings being overlooked, and ultimately inaccurate conclusions being drawn, with an unfavorable effect on business decisions, actions, and outcomes. Underlining this is Forrester’s research that determined only 29% of organizations are successful at correlating analytics to actions.

While the challenges organizations face are significant, the solution to many of the roadblocks can be found in augmented analytics. A compare of data management, data science, and data visualization depicts the benefits that augmented analytics delivers.

Data Management	Data Science	Data Visualisation
Analytics Workflow – Current State
45% of time is spent handling manual tasks such as data cleaning, profiling, cataloging, etc.	34% of time is spent on manual feature engineering model selection, and model training and deployment	21% of time is spent using interactive techniques such as filtering, pivoting, linking, grouping, and user-defined calculation
Issue: Less time spent on more productive tasks such as gaining insights from the data	Issue: Less time spent on model validation, testing delivery, and operationalization	Issue: Relies on the individual user to interpret the data
Augmented Analytics Workflow
Data preparation and discovery time is reduced by 50 – 80%	Automates data science tasks, reducing time spent by 40%	From query to insights, cycle time is reduced by 50%
Issue: Manual data preparation, data quality, and cataloging	Issue: Manual feature engineering and model building	Issue: Manual exploration of data using interactive visualisation
Solution: Uses AI to automate data preparation	Solution: Uses AI/ML techniques (AutoML – an AI-based solution, automates the process of applying ML to real-world problems) to automate data science tasks such as auto generation of features, and model management is augmented	Solution: Uses natural language processing (NLP) for auto visualisation of relevant patterns, and automates data insights
Benefits: Increases productivity and efficiency	Benefits: Improves accuracy of the model and removes bias	Benefit: Provides faster insights from the data

Augmented analytics democratizes AI across the data value chain, automating the data preparation process and key aspects of data science, while natural language processing (NLP) enables users to obtain faster insights. AutoML leverages techniques and relevant insights using NLP and conversational analytics, and includes:

• Augmented data preparation: Uses AI and ML to accelerate manual data preparation tasks, such as data profiling and quality, enrichment, metadata development, data cataloging, and various aspects of data management, including data integration and database administration.
• Augmented data science: Uses AI and ML techniques to automate key aspects of data science, such as feature engineering and model selection, as well as model operationalization, model explanation, and model tuning.
• Augmented analytics: A component of BI platforms, it embeds AI/ML techniques to automatically find and visualize the data and narrate the relevant findings through conversational interfaces, including natural language query (NLQ) technologies that are supported by natural language generation (NLG).

Realizing the benefits of augmented analytics

At a high level, the benefits of augmented analytics include automating data preparation, reducing time to insights, eliminating human analytical bias, and mitigating the risk of missing important insights. In addition, augmented analytics provide the ability to democratize data analytics for less business-savvy users, such as citizen data scientists that don’t have the specialized training or skills in data science or analysis.

The benefits of augmented analytics can’t be disputed. As the amount of data generated by an increasingly connected world continues to rise, organizations will more than ever require the capabilities augmented analytics provides. They will need to connect disparate and live data sources, find relationships within the data, create visualizations, and enable personnel to share their findings quickly and effortlessly. The use of augmented analytics is poised to change how users experience analytics and BI by providing insights that are currently unattainable.

About the author:
Rohit is responsible for delivering business growth through innovation and product strategy. He leverages his expertise in Artificial Intelligence (AI), analytics and digital services to contribute to Subex’s solutions. Prior to this, Rohit was the Head of Subex’s Business and Solutions Consulting Group (BSCG) and has over 20 years of rich experience in delivering solutions and consulting to telecom operators across the world. Before joining Subex, Rohit worked with companies like Crompton Greaves and Kirloskar Electric Company. He is a graduate in electrical and electronics engineering from University of Mysore.

 

 

If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!

Categories: Big Data · Industry Viewpoint