{"id":30,"date":"2022-06-09T12:52:33","date_gmt":"2022-06-09T12:52:13","guid":{"rendered":"https:\/\/eck.uidev.site\/?page_id=30"},"modified":"2023-07-14T17:13:34","modified_gmt":"2023-07-14T17:13:34","slug":"takeaways","status":"publish","type":"page","link":"https:\/\/markets.eckerson.com\/next-generation-data-catalogs\/takeaways\/","title":{"rendered":"Key Takeaways"},"content":{"rendered":"
\n\t
\n\t\t\t\t

Key Takeaways<\/h1>\n\t\t\n\t\t
\n\n\t\t\t\t\t\t
\n\t\t\t

Wayne Eckerson and Kevin Petrie summarized the key points heard during the CDO TechVent session. Here is a summary of their conversation:\u00a0<\/span><\/p>\n

    \n
  • Wayne\u2019 s \u201cAge of Metadata\u201d seemed to resonate with the audience and speakers.\u00a0<\/span><\/li>\n
  • Audience member Bob Seiner commented that although a data catalog automates metadata collection, it\u2019s important to spend time with business people to understand how they\u2019re using the data catalog. Can\u2019t replace the people element!\u00a0<\/span><\/li>\n
  • There was a lot of discussion around tailoring a data catalog to individual personas. A single catalog can support multiple personas if access controls are applied correctly. This made Wayne quip:\u00a0 \u201cIt\u2019s important to person-ify the data catalog\u201d<\/span><\/li>\n
  • Lewis Wynne-Jones mentioned the \u201cImmaculate GUI\u201d which amused Kevin and Wayne who also agreed with the point of having an easy-to-use GUI.\u00a0<\/span><\/li>\n
  • Kevin was taken aback by the percentage of people with \u201cNo plans to implement a data catalog\u201d in the survey data. Wayne responded that this creates an opening for mid-market products that are affordable, fast to deploy, and easy to use.\u00a0<\/span><\/li>\n
  • There was quite a bit of discussion about the impact of large language models on the data catalog market. Here\u2019s the summary:\u00a0<\/span>\n
      \n
    • LLMs are good for giving users a \u201cwarm start\u201d as Kaycee Lai said\u00a0<\/span><\/li>\n
    • They can automatically create table descriptions, field labels, glossary descriptions<\/span><\/li>\n
    • But someone still needs to check for accuracy.\u00a0<\/span><\/li>\n
    • They can help with natural language search to improve data discovery\u00a0<\/span><\/li>\n
    • LLMs pose a security risk if pointed at actual data. LLMs are thus best used for metadata not for data.\u00a0<\/span><\/li>\n
    • Kevin mentioned the rise of Small Language Models that model domain specific text and thus deliver more accurate results.<\/span><\/li>\n<\/ul>\n<\/li>\n
    • There was also lots of discussion around data virtualization, which all three of our sponsors support. The feeling was \u201cWhat\u2019s the point of providing access to metadata, if you can\u2019t access the data.\u201d We couldn\u2019t agree more! <\/span><\/li>\n<\/ul>\n\t\t\t<\/div>\n\t\t\t\t\t\t
      \n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\n\n\t\t\t\t\n\n\n\n\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/div>\t\n\t<\/div>\n<\/div>\t\n\n
      \n\t
      \n\t\t\n\t\t
      \n\n\t\t\t\t\t\t
      \n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t
      \n\t\t\t\t\t

      Some speakers joked that we’ll be discussing metadata swamps in three years!<\/p>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\n\n\t\t\t\t\n\n\n\n\n\t\t\t\t\t\t\t\t

      \n\t\t\t\t\t
      \n\t\t\t\t\t\t\t\t\t\t\t
      youtu.be\/8pzLHIOcDOk<\/div>\n\t\t\t\t\t\t