<< back Page 2 of 4 next >>

The Limitless Applications of Analytics

Furthermore, the application of analytics to the pandemic is not just for commerce. Government agencies can rely on analytics to recommend and set policies based on reliable, data-based predictions. And school districts can use analytics to better plan for when schools should be shut down, reopened, or conducted online.

AI-Augmented Analytics

Perhaps the most significant trend impacting data and analytics is the integration of AI into the process, commonly referred to as augmented analytics. This was first identified by Gartner and defined as using enabling technologies such as AI and machine learning (ML) to bolster data preparation as well as the generation and explanation of insights derived from data.

Organizations are increasingly deploying augmented analytics to improve the description and connection of multiple internal and external data sources and to find relationships within the data. The difference is that augmented analytics automates many aspects of the process, thereby making it possible for non-experts to perform work such as developing, managing, and deploying models, which previously required skilled data scientists.

Of course, as is the case with most AI capabilities, adoption of augmented analytics is nascent, but growing. As data volumes continue to expand, automating analytics with AI will continue to escalate in importance as it becomes impossible to manually deal with the incredible volume of data.

Using AI and ML to augment data preparation is another aspect of augmented analytics adoption that deserves to be mentioned. The manual tasks required to prepare data before it can be modeled or analyzed are some of the most time-consuming aspects of data analytics and data science. Automating data preparation with AI and ML to profile and enrich the data, define metadata, catalog the data, and perform the other mundane aspects of getting data ready for analysis is being used to accelerate analytics processes. As these AI and ML capabilities improve over time, expect to see a reduction in the amount of upfront planning and preparation time required. In addition, data quality may be improved also as the AI/ML techniques improve.

Obviously, the problem of integrating AI into analytics has not been completely solved. Many issues remain, such as a lack of skilled technicians, the cost of adoption, software immaturity, and the length and difficulty of pilot projects. Nevertheless, AI is burgeoning and will continue to be embraced as augmented analytics improves and succeeds.

As more AI capabilities are embedded into hardware, the velocity of augmented analytics adoption will increase. One such effort to keep an eye on is Telum, IBM’s latest processor for its IBM Z mainframe computer. Telum contains on-chip acceleration for AI inferencing while a transaction is taking place. Announced in August 2021, the first Telum-based mainframe is expected to be available in the first half of 2022.

Adoption of Data Fabric Architecture

A data fabric architecture is deployed to simplify and integrate data management in a hybrid environment spanning both on-prem and cloud implementations. The data fabric integrates data and connecting processes using analytics over a set of data services to provide consistent capabilities for accessing and using data across the entire hybrid environment.

Most organizations rely on a myriad of operational systems, both on-prem and in the cloud, that are based on multiple technologies and requirements. The data fabric confers the following benefits on the analytics processes that take advantage of it:

  • Access to the business, operational, technical, and social metadata that defines the data through the data fabric, as well as the continuous analysis, modification, and depiction of metadata assets.
  • Continuous learning with predictions for data management and integration requirements and capabilities.
  • The ability to create and curate knowledge graphs that enrich data with semantics to make it more intuitive for data analytics usage.
  • Built-in compatibility for multiple data integration techniques, such as ETL, replication, streaming, and so on.
<< back Page 2 of 4 next >>


Subscribe to Big Data Quarterly E-Edition