Most data warehousing projects are riddled with tech debt and bad design, which has led to IT leaders avoiding classic data warehousing approaches.
But there are modern tools and approaches that take the pain out of ETL (extract, transform, load) and data warehousing, says Evan Barke, senior consultant at Keyrus. "These tools can leverage and include the modern data systems like big data and machine learning."
Barke will speak at ITWeb Business Intelligence & Analytics Summit 2019, to be held on 12 and 14 March, at The Forum in Bryanston.
Migrating to a modern data warehousing architecture has its challenges, but it brings significant benefits, too. One of the difficulties, notes Barke, is establishing how to handle big data and machine learning in a modern hybrid BI system.
"There is no panacea that handles all BI architecture requirements, but there are ways of pulling the different systems together and having the best of all worlds if clever choices are made.
"A properly built modern architecture allows businesses to leverage valuable information from modern big data platforms, IOT and machine learning, while keeping the valuable structure and ease of access of classic star schema systems," he says.
During his talk, Barke will discuss the extent to which existing data warehouses can provide support for data science, self-service and advanced analytics. He will also examine the new technologies that are superseding the traditional data warehouse, such as logical data warehouses, data lakes, distribution hubs, data catalogues, analytical sandboxes and data science hubs.
He will give data practitioners and warehouse managers advice on how to create balance between the dependability of mature data warehousing technologies and the advantages the newer technologies provide.
Share