By 2025, seventy percent of organizations will focus on small and wide data instead of big data. According to forecaster Gartner, the turnaround is due to users trying to put their analyzes in a deeper context, while also wanting to diet-hungry artificial intelligence.
At a time of subversive changes, such as COVID-19 proven, historical data reflecting the past is rapidly becoming obsolete, rendering many models of artificial intelligence (AI) and machine learning (ML) used in live operations unusable, said Jim Hare, vice president of research at Gartner. In addition, AI support for human decision-making has now become complex and resource-intensive because it relies too heavily on data-hungry methods of deep learning.
D&A (data and analytics) managers responsible for managing and utilizing corporate data assets therefore, they are turning to new analytical techniques that have become known as small data and wide data. By using the two methods together, organizations can use their data more efficiently, the analyst pointed out, either by reducing the amount needed to teach AI models or by gaining more valuable insights from unstructured, diverse data sources.
More robust analytics and AI
Small data is an approach that provides useful insights from less data, including time-series analytical techniques or few-shot learning, synthetic data, or self-monitoring. supervised) uses learning. And broad data, which allows the aggregation and joint analysis of different small and large, unstructured and structured data sources, uses so-called X-analytics to explore the relationships between data sources and to manage a variety of data formats, including tabular, textual, visual, audiovisual, sound, temperature, or even odor detection and vibration sensor data.
Both approaches can increase the robustness, ie reliability, of analytics and artificial intelligence, leaving the organization less reliant on big data, yet it gains fuller situational awareness, 360-degree visibility, and can make better decisions based on analysis, Jim Hare said. Data and analytics managers can apply both techniques with success if they have less data to teach AI or build more robust models using a variety of data.
Small and wide data could have potential applications in commerce, for example. forecasting or real-time behavior and mood analysis in customer service, customizing services, and improving the customer experience. The two methods are also well-suited for physical (facility) protection, fraud detection, and adaptive, self-controlled systems such as robots, which continuously learn from the temporal and spatial relationships of events by analyzing data from their sensors.
Data Tissue and the Endowed Consumer
In its forecast, Gartner highlighted small and wide data focus. The technologies and methods on the list help organizations respond to change, manage market uncertainty, and take advantage of new business opportunities. D&A managers should therefore also study the following nine trends as potential areas for business-critical investment.
Smarter, more responsible and scalable AI – With artificial intelligence and machine learning, companies can achieve greater business impact if they are smarter, they use new techniques that result in less data-hungry, ethical, responsible, and more resilient, more reliable AI solutions. With such solutions, they can build faster learning algorithms and meaningful systems that can create more value in less time.
Composable data and analytics environment
– Open and containerized analytics architectures make it easier and more flexible to build the analytics capabilities you want. From a variety of data sources, analytics tools, and AI solutions, organizations can quickly build user-friendly, intelligent applications that combine analytical insights with action. As the center of gravity of the data universe moves into the cloud, computable analytics becomes an agile method of application building supported by cloud-based marketplaces, low-code, and no-code tools.
Foundation Data Tissue (data fabric) – The impact of digitalization is becoming more and more felt, users are becoming more and more capable and independent, therefore more and more R&D managers are dealing with the challenges of diversity, distribution, scaling and complexity with data fabric. in its data environment. The data fabric uses analytics to continuously monitor data usage, dynamically align components across hybrid and multi-cloud environments, and reduce design and deployment lead times by 30 to 30 percent and maintenance time by 70 percent by combining and recycling different data integration styles.
XOps, efficiency in everything – Purpose of the XOps approach (including DataOps, MLOps, ModelOps and PlatformOps) ) to increase efficiency and economy in analytics through recycling and repeatability, along with DevOps good practice, along with reliability. By avoiding duplication of technologies and processes, XOps also promotes automation in the data environment. Many analytics and AI projects run aground because companies would address operational efficiency issues retrospectively after development. With XOps, D&A managers can avoid this pitfall by extending the repeatability, traceability, integrity, and connectivity of analytics and AI tools to the enterprise level.
Decision Intelligence Engineering Planning – Instead of supporting one decision at a time, this planning focuses on a series of decisions that are grouped and integrated into business processes, or even organized into networks of evolving decisions and consequences. Decisions are becoming increasingly automated, so by planning the process more extensively, D&A managers can increase the accuracy, repeatability, transparency, and traceability of decision-making.
Data and analytics as a business basic function – All activities related to data and analytics in companies are increasingly becoming a basic business function from secondary occupation. Better collaboration between central and distributed D&A teams breaks down existing technology silos, sharing assets across the organization so you can put them to business performance more effectively.
Everything connecting graph – Modern data and analytical capabilities are based on graphs that also explore the relationships between people, places, things, and events through diverse datasets. With the help of graphs, D&A managers can also quickly answer complex business questions that require knowledge of the context and an understanding of the nature of the relationships. Gartner predicts that 80 percent of data and analytical innovations will have graph technologies to emerge by 2025, promising a leap forward from 10 percent this year and giving a huge boost to organizational decision-making.
The Authorized Info Consumer – Most business users today still use pre-built digital dashboards or manually discover data, which can lead to erroneous conclusions and bad decisions. However, current practice will be gradually supplanted by automated, interactive analytics available in a dialog-based interface that performs dynamically generated analyzes according to the individual needs of the user and delivers the findings to the place of consumption. As a result of this trend, the consumer of information is increasingly acquiring analytical skills that were previously only available to professional analysts and citizen data scientists.
Data and analytics are on the edge network – An increasing number of data, analytical and other technologies supporting the operation of organizations are migrating to edge computing, near devices operating in the physical world, outside the competence of the IT department. According to Gartner, in 2023, more than half of the primary tasks of D&A managers will already cover data that is generated in a perimeter network environment and must be managed and analyzed there according to the needs of real-time and automation. To do this, data monitoring and management must become more flexible and responsive, but also more resilient
Hardware, software, tests, curiosities and colorful news from the IT world by clicking here
