Data vault modeling in a complex data environment - Hendrik Ruijter & Anders Lindström

Поділитися
Вставка
  • Опубліковано 30 вер 2024
  • We want to share our experiences in implementing a cloud-based Data Warehouse data vault model. Our company’s operational systems collect telemetry data, real time events and surveillance data from millions of devices. We have all the corporate systems related to logistics, finance and other data storage solutions used in different parts of the organization. A data vault model is meant to be robust towards changes in source systems, data should be loaded in parallel, and additions to the model should be easy to integrate. Here be dragons.
    Hendrik Ruijter has 20+ years’ experience from e.g. Axis, SonyEricsson and Sony in different roles. He works with cloud data warehouse development in Verisure Innovation. He is interested in data vault modeling as a bridge between operational 3NF models and dimensional report models. Software-as-a-service data warehouses like Snowflake is another interest.
    Anders Lindström has 10+ years’ experience as a consultant within Data Warehousing and Business Intelligence. He works at Combitech helping customers like Verisure to realize their dream of a common data platform built for analytics. For the latest years, he has been working primarily with Data Vault modeling in order to solve integration issues in a business-friendly way. Data Warehouse automation is another interest.
    _____________________
    Foo Café is a meeting place for people with an interest in software development and IT.
    Here, user groups and partner companies find the perfect atmosphere for sharing knowledge and develop competence.
    Foo Café is open for everyone and is free to attend. To make an event at Foo Café, you have to become a partner company with us or you can run a user group with an in-depth technical focus.
    We welcome companies with sponsor challenges.
    We are currently present in Stockholm and Malmö and have made over 2048 tech events since the start in 2012.
    At Foo Café there’s an event going on everyday, year round.

КОМЕНТАРІ • 4

  • @agustinespinosa4096
    @agustinespinosa4096 4 роки тому +1

    58:19 Business analysis and SMEs - Very Interesting and important

  • @pecheur1234
    @pecheur1234 5 років тому

    Hi, I can not read the video, it stops at 8:47. In addition, English subtitles are not available. Can you do something ? thank you !!

  • @tarpara
    @tarpara 3 роки тому

    Why is “Steering_Wheel” a hub and on a satellite of Car_Hub?

    • @hendrikruijter2151
      @hendrikruijter2151 3 роки тому

      One option is to say that all cars have one steering wheel. A car will never have two or more steering wheels. You would argue that context close to key makes the steering wheel a SAT of the car HUB. This SAT would keep all attribute information of steering wheel: does it have paddles, does it have a Siri button, etc. If you believe in multi-valued satellites you can add information for a second steering wheel in the specification of the car. Another option is to handle uncertainty when you are not a car expert, maybe there are cars produced with more than one steering wheel in a car. When the link creates a relation between a car HUB and a steering wheel HUB you can say that the steering wheel is a core business concept. The business talks about steering wheels. Steering wheels with many options sell cars to customers. Now, the steering wheel SAT is related to the CBC of a steering wheel HUB and all the SAT attributes are the same as in the context close to key SAT.Adding a second steering wheel would require a linked key model choice. The basic link with one car HUB and one steering wheel HUB is like a faceless fact table. You know that one car has one steering wheel. To model the relation you must add more information.