Demystifying Augmented Analytics—Already?

In Data Governance, Data Management by IRM UKLeave a Comment

Print Friendly, PDF & Email

“The world of the future will be an even more demanding struggle against the limitations of our intelligence, not a comfortable hammock in which we can lie down to be waited upon by our robot slaves.”
– Norbert Wiener, American Mathematician and Historian

The one thing I can’t complain about in my beloved market of BI and analytics software is the lack of new terms and approaches and incorporation of new technology. In fact, it is a fast-moving and -evolving market.

Jorge García, Principal Analyst, TEC
Technology Evaluation Centers (TEC) is a global consulting and advisory firm ; they are Media Partners of the Data Governance Conference & Master Data Management Summit Europe 13-16 May 2019

This is especially true when you consider that in recent years, more and more organizations have been looking for ways to increase their arsenal of data management and analytics capabilities—and use robust data analysis technologies for achieving optimal business performance and operations.

A little more than a year ago, Gartner published a report called, “Augmented Analytics Is the Future of Data and Analytics,” coining a new concept of “augmented analytics.” Some of my readers and customers ask questions such as, “What is this new concept and what does it entail?” “What are the benefits and challenges of augmented analytics?” and, “What are the misconceptions of this new type of analytics paradigm?”

In what follows I provide my views of this new trend of augmented analytics.

What Is Augment Analytics?

In the introduction of the report, Gartner refers to augmented analytics as:

“an approach that automates insights using machine learning and natural-language generation”

that

“marks the next wave of disruption in the data and analytics market.”

According to the same report, the augmented analytics approach focuses on engaging machine learning to “augment” or increase efficiency on three main aspects of the analytics process:

  1. Data Preparation. Applying machine learning to data preparation is one of the most labor-intensive aspects of analytics. It aims to make it easier for users to clean, prepare, and optimize the formatting of data for analytics purposes. By incorporating machine learning, new analytics software solutions can provide users with special features to automatically detect and impute (replace) missing values, automate date conversion and text replacement, or import external calendars.
  2. Data Discovery. Incorporating machine learning within the data discovery functionality features of software can help users to avoid having to create models in order to gain insights and improve visualization. It also allows them to automatically create histories and narrations.
  3. Data Science and Machine Learning. Incorporating machine learning into the data science process supports the automation of different aspects of advanced analytics, from the generation to the management and operation of models, giving less-knowledgeable users access to these types of solutions while enabling savvy users to save time with performing such tasks.

By infusing machine learning into these key aspects of analytics software, software providers aim to empower their users with tools to improve data analysis services—by enabling automation of the process and a cycle of continuous improvement.

What Are the Most Common Augmented Analytics Capabilities?

By infusing machine learning capabilities into their data management and analytics software solutions, providers of new augmented analytics solutions aim to favor automation and simplification in order to reduce the time and complexity of data analytics processes. Below are the major functionality features and the specific capabilities of each:

Augmented Data Preparation

  • Automated data inference that intelligently infers semantics in the data.
  • Automated data harmonization that identifies relationships across all data dimensions.

Augmented Data Discovery

  • Automated workflow analysis
  • Enhanced recommendation on how data should/could be arranged, mashed up, and visualized

Augmented Data Science

  • Automated machine-learning feature selection
  • Automated testing to select the optimal model

With these features, automation and simplification play a big role in applying machine learning and AI techniques to the typical analytics workflow.

What Are the Main Benefits of Augmented Analytics?

“Augmented” analytics software solutions bring more ease of use and enhanced automation capabilities to the analytics process cycle. As such, they can provide general benefits to the overall BI and analytics process. They can:

  • Shorten the entire analytics process cycle.
  • Automate many key analytics tasks.
  • Widen (democratize) the user footprint.
  • Allow more information workers to benefit from the use of analytics services.

Consequently, these solutions may lead to technical benefits for information workers, which can translate into benefits for the business:

  • Due to ease and automation, data expert teams can concentrate on strategic business goals and data scientists on the methods to achieve them.
  • More and more complex analytics practices can become routine analytics practices, reducing the time to perform analytics and favoring data and insight consumption.
  • More and more analytics flows will be able to be repeated across different areas of the business.
  • Automation can provide improvements in data management technologies that promote user adoption.

Is Augmented Analytics the Panacea to All Our Problems?

The answer to that question is most likely no. Why is that?

Despite their potential to bring enormous benefits to most, if not all, organizations, augmented analytics software solutions require that organizations put into practice key principles of information technology (IT) management. So, regardless of their method of implementation, these solutions are concerned with data quality, insight provision, and data governance:

  • Data Quality. This is governed by the garbage in garbage out (GIGO) iron rule of computer science, where the quality of the data input still has a direct effect on the quality of the data output. With the addition of machine learning to the mix, the logic of the rules has become fixed and rigid, ensuring that the data used with models is of high enough quality to ensure a high-quality outcome. While the preparation of augmented data may entail a time-consuming process, it is still necessary to ensure data quality is optimal for an optimal outcome. As I read from a presentation on Strata on AI and Data Governance:

“3 trillion images of cats will not inform me about the shape of a dog.”
– Jean-Francois Gagne, CEO of Element AI

  • Data and Insight Dissemination. As machine learning occupies a more and more important role in the decision-making process of the organization, the potential for technology to alleviate the role of humans in making decisions about the business, or remove them altogether, increases as well. This potential marginalization of the role of humans in business decision-making changes the way data governance has traditionally operated. Today, it is not just about the governance process (manipulation and movement) but about understanding even more about the data. This includes everything from data collection, to the correct use of the dataset to train an AI model to understand biases and maintain traceability, accuracy, accessibility, and privacy. Organizations will find that there’s still much work to be done to ensure data is provided to users according to their rights and privileges and in the format they require for specific corresponding purposes.

“An updated notion of genius would have to center around
 one’s mastery of information and its dissemination.”

– Kenneth Goldsmith, American Poet and Critic

  • Data Governance. Traditional approaches have focused mostly on human-based processes. That is, who, when, and how people saw, see, and will see the data. They have not centered around the “automatic” provision of insights and data from the system. There’s currently a need for a balancing act to be able to ensure access and agile data management and provision for humans, as the ultimate goal of any data governance initiative is to ensure business, life, or health improvement through knowledge.

“The goal is to turn data into information, and information into insight.”
– Carly Fiorina, American Businessman (former CEO of Hewlett-Packard)

There are many advantages to having augmented analytics capabilities available. Implementing augmented analytics capabilities within your organization will empower business users and information workers with the knowledge they need to make more informed and accountable decisions.

Still, to maximize the chances that any augmented analytics initiative will succeed requires that the initiative be embedded within a larger and comprehensive data management strategy.

Are There Alternatives to an Augmented Analytics Approach?

The software industry is replete with new or alternative approaches to improving business processes, with augmented analytics serving as one such approach. Software providers and organizations frequently introduce new methods and models for conducting analytics.

There are two aspects of augmented analytics that aim to help users make more reliable business decisions:

  • Data Intelligence. This approach aims to provide access to a wide variety of analytics capabilities such as descriptive, predictive, and prescriptive analytics. It also includes autonomous decision-making, machine learning functions, and visualization tools to provide both savvy and novice users access to high-quality analytics services that can be deployed across public clouds, on premises, on an optimized or commodity infrastructure. Data intelligence focuses on the analysis and interaction with information in a meaningful way to promote better decision-making.
  • Smart Data Discovery. This approach aims to provide next-generation data discovery capabilities. It provides insights generated from advanced analytics capabilities that can be easily translated to business users or citizen data scientists, without requiring them to have traditional data scientist expertise.

These two methods incorporate advanced analytics capabilities for achieving extreme ease of use. Data intelligence and smart data discovery can be seen as two “new” approaches that can complement an ongoing augmented analytics strategy, or even serve in place of one.

Companies, particularly those that have an augmented analytics implementation project in progress or on their agenda, should explore the potential of these two approaches.

A Look to the Future

Whether augmented analytics has the potential to trigger a full transformation of the BI and analytics market in the coming years remains to be seen. At the moment, it is safe to say that augmented analytics is part of a global effort to advance analytics and BI to a next level of service. It enables companies to transform and expedite their common analytics processes and achieve the complete flow of data from source to target.

Consequently, one can expect that new BI and analytics offerings will increasingly incorporate automation features within all analytics processes, providing automated data preparation and data visualization. Therefore, we can expect traditional BI solutions to move more and more into the realm of data science and to merge with data science platforms.

With the incorporation of augmented analytics and other high-end analytics technologies, organizations have the opportunity crystalize important benefits.

Businesses can consider having augmented analytics within their corporate software stack to allow information workers from both IT and business areas to focus more on strategic initiatives and special projects, thereby positively impacting the organization’s return of investment (ROI) and total cost of ownership (TCO) of the overall analytics portfolio.

Yet, to make sure your organization benefits the most from these technologies, a new organizational mindset is required, based on your organizational needs. For this, you could consider some guidelines:

  • Data is a key part of an organization’s digital assets and, as such, can represent a competitive advantage. So before engaging in a new analytics strategy, take some time to reflect on how data and digital assets will potentially become a competitive advantage against your competitors.
  • Learning to scale requires that leaders and information workers first apply such new analytics capabilities to real problems or case studies and then evaluate their efficacy, scaling them up only when these capabilities have been proven effective.
  • Data analytics and decision-making processes will need to change and adapt as new analytics and AI capabilities get introduced into the market.
  • Interactions between humans and machines will change, so that a new approach to human capital and acquisition and development might need to be considered, especially regarding soft skills (business, management, analytics, etc.).

Providers need to be able to provide an organization with the right software tools for dealing with the processing and analysis of increasing volumes of data today and in the years to come. Having these augmented analytics capabilities will become increasingly crucial to ensuring business improvement and success.

Augmented analytics could be your next best partner for this success, no matter what role you play within your organization. For more news on BI and analytics, refer to the TEC Blog.

Jorge Garcia is the principal analyst of business intelligence (BI) and data management at TEC. With over 20 years’ industry expertise, Jorge also specializes in manufacturing, project and process management and business process management.

Copyright Jorge García, Principal Analyst, TEC

Leave a Comment