These benefits make it easier to access data that exists in multiple silos throughout the organization. As such, deploying a data fabric architecture can help to alleviate vexing problems such as poor data quality, missing data, and inaccessible data. Building upon a single data fabric eliminates problems that arise from using multiple and different types of database systems with disparate data management tools.
Although adoption of a data fabric architecture is only just emerging, prescient organizations are deploying data fabric architectures to modernize data management and data integration for analytics. Organizations that store copies of data sourced from multiple (often thousands) of systems can benefit greatly from the data fabric architecture.
Marriage of the Cloud and Analytics
Less of a current trend and more of an ongoing one, deploying analytics in the cloud continues to gain traction. Cloud analytics—or carrying out analysis using cloud computing—requires a range of analytical tools and techniques to help companies extract information from big data and present it for consumption, typically via a web browser.
Until recently, cloud analytics was generally considered to be analytics performed exclusively on cloud data, but this is no longer the case. For medium-to-large enterprises, the trend is toward hybrid analytics, where cloud analytics solutions (those that run in the cloud) can be used to analyze any data, whether it is on-prem, in the cloud, or both. Analytics services and tools are expected to be able to satisfy such scenarios, and usage of those that cannot is waning.
Adoption of Graph Technology and Techniques
The adoption and usage of graph techniques and database systems is accelerating because of their strong ability to map and analyze relationships. Graph theory is based on the mathematical graph structure, which is used to model relations between objects.
When massive amounts of data are stored throughout an organization, in the cloud, and distributed across multiple environments, the use of graph techniques makes it possible to discover and manage multiple connections between data sources. Furthermore, when organizations need to perform analytics primarily on relationships between data, rather than on the data itself, graph technology can be invaluable.
The Economics of Analytics
Organizations are increasingly implementing data analytics projects with the goal of understanding, measuring, and monetizing these projects instead of treating analytics as experimentation. The actionable insights achieved by data analytics projects can contribute to the bottom line of an organization in many significant ways.
By applying economic principles to the deployment of information assets, the insights from analytics can be used to improve innovation, make better decisions, generate increased value, and thereby improve organizational success.
Clearly, a comprehensive approach to the economics of analytics requires organizational effort and buy-in, a firm understanding of both economic and data management principles, and much more. For an in-depth study of such efforts, consult the recent books, Infonomics by Douglas Ladley and The Economics of Data, Analytics, and Digital Transformation by Bill Schmarzo.
As part of this effort, organizations are shifting data and analytics to become a core business function, instead of a component of IT that supports the business. The long-awaited shift to treating data as a business asset appears to be closer to becoming a reality than it ever has been.
Edge Analytics and DataOps
The growth in the amount of data being created and stored is a trend that has continued unabated for some time now. And with the increasing number of sensors being connected to IoT, more data is being generated and processed every day.
The growth of edge analytics has resulted from this confluence of trends. Edge analytics refers to the process of gathering, analyzing, and creating insight directly from the devices connected to IoT, instead of passing the data from the device to a central computing environment. Edge analytics goes one step beyond edge computing by performing the analytics on the data where it is captured.
Yet another trend resulting from data growth is that more organizations are deploying DataOps, a methodology for automating data management to improve data quality and reduce the time needed for preparing and utilizing data for analytics. DataOps is not just DevOps for data. DevOps is a philosophy and set of practices that integrate development and operations with the goal of improving software delivery and quality. DataOps encompasses practices, workflows, procedures, and cultural norms to bolster and improve the lifecycle of data. DataOps requires collaboration between and among all pertinent IT professionals, from data architects to DBAs and data scientists to developers. It improves all aspects of data management through processes such as data creation and ingestion, data quality steps, data preparation, data accessibility, reporting, and beyond. As the amount of data continues to grow, DataOps is becoming a crucial component of being able to glean insight from analytics.