Jimmy Pang

Form

Subtitle

Jimmy Pang

From

Subtitle

Why Castor

Castor has been around a year. Very nice tool for him as a BI tool. One of the biggest challenge for Vestiaire as a BI team: the management has good data literacy and can conduct their own analysis, they need to prepare the data in a reusable format. Very important to address how the information flows within the organisation. When you are 100 people, you can tap on the shoulder to get information. After that, these kind of practices are outdated, inefficient and dangerous. You end up addressing ad hoc request 100% of your time. Castor scans the data stack (Tableau, Snowflake, DBT). Auto-sync is very useful because it helps to have an overview in a user friendly manner to explain the data stack. A useful feature is ownership. You need to know you to talk to and who is responsible. Feature request (data export). Very powerful for cloud migration. Very amazing feature: overview of what happens in Castor (castor internal usage) + data stack (how many table and view we have per stack?). Pin point the quality of the documentation.

What do you do on Castor?

Trying to push his team to document. Every single ticket will ask people to document data. Castor is meant to be a catalog tool, it feels that if Castor does too many things it’s a problem. Firm believe of eating your own dog food. He is using his own documentation. Between IT and BI isn’t always easy. Discovery can be solved by Castor but a lot of work has to be integrated. Lineage is very useful.

value

Overview

  • Vital. Overview of data. Metadata from Castor. Data about data.
  • If Castor want to push this to a higher level, it would be to export metadata easily from the app.

Castor homepage:

  • top table are useful.
  • Recent contributors useless. Not sure what actions he should take from that.
  • Latest comments. Not always useful. Not sure communication should happen on Castor.

Feature Request:

  • Not only limited to a certain tech stack. Bringing all data assets from various tech stack together.
  • Usage of data assets. Help bring visibility on who’s using what. What should we maintain and improve based on usage?
  • Dependencies for migration/maintenance.