Modern Data Integration for Advanced Analytics, from Self-Service to Predictive Solutions
Webinar Speaker: Philip Russom, Senior Research Director for Data Management
Date: Tuesday, February 25, 2020
Time: 10:00 a.m. GMT
We say “analytics” as if it is one monolithic thing. In actuality, there are many analytics techniques, practices, and enabling technologies. These include older forms such as online analytical processing, statistics, performance management, and self-service practices, as well as newer and more advanced forms such as data mining, text mining, natural language processing (NLP), clustering, and graph. At the moment, the hottest growth is seen in machine learning and artificial intelligence as enablers of predictive analytics.
The point is that each form of analytics has its own unique set of data requirements, which in turn affects how data should be acquired, integrated, and transformed on a per analytics approach or use case basis. For example, OLAP works best with data transformed into dimensional models, whereas data mining works best with unaltered source data. Similarly, self-service data exploration and prep works well with lightly standardized structured data, whereas text mining and NLP are optimized for unstructured data. Hence, data’s journey into analytics solutions varies considerably.
Attendees will learn about:
- How and why the increasing business use of advanced analytics is the leading driver for change across data management, with special considerations for data integration
- How cloud-based data platforms, databases, and tools—and hybrid architectures—support multiple forms of analytics and their requirements for data storage and integration
- The highly diverse data requirements of ten or so analytics approaches, with a focus on how this affects tool selection and solution design for data integration
- Detailed dives into modern data integration and data prep for popular analytics, such as self-service data practices, data visualization, and NLP
- The especially complex data requirements of machine learning and analytic model development, plus how data integration and data prep should address these
Philip Russom, Ph.D.