Organizing Dataiku Projects & Folders For Multi Departments Usage

Solved!
lina
Level 2
Organizing Dataiku Projects & Folders For Multi Departments Usage

Hi, 

What are the approaches you follow and the best practices for organizing and structuring Dataiku projects and folders for the use of 60+ employees from different departments and business functions? 

Best regards, 

0 Kudos
1 Solution
Manuel
Dataiker Alumni

Hi Lina,

It is hard to give specific advice, because it really depends on your projects and your organisation structure.

Perhaps a point that you are missing is that the displayed projects are primarily governed by the security framework, i.e., a user only sees the projects they have access to. So, before thinking of folders, you should think who has access to which projects:

  • At Dataiku, in one of our DSS instances, the security groups match our roles at Dataiku (customer_sucess, data_science, etc.), so depending on your role, you will only have access to the projects relevant to your role;
  • At some customers, security groups are organised around access to data sources, so a user only sees projects with data they can see
  • Every customer has to tailor the security framework to their specific needs

You should also review the Workspaces capability, a recent feature with v10. It is a way to organise related assets from multiple projects: 

Finally, other best practices such as naming conventions and documentation are also useful to get projects organised. Some suggestions on this page: https://knowledge.dataiku.com/latest/courses/governance/collaboration/best-practices-collaboration.h....

I hope this helps.

View solution in original post

1 Reply
Manuel
Dataiker Alumni

Hi Lina,

It is hard to give specific advice, because it really depends on your projects and your organisation structure.

Perhaps a point that you are missing is that the displayed projects are primarily governed by the security framework, i.e., a user only sees the projects they have access to. So, before thinking of folders, you should think who has access to which projects:

  • At Dataiku, in one of our DSS instances, the security groups match our roles at Dataiku (customer_sucess, data_science, etc.), so depending on your role, you will only have access to the projects relevant to your role;
  • At some customers, security groups are organised around access to data sources, so a user only sees projects with data they can see
  • Every customer has to tailor the security framework to their specific needs

You should also review the Workspaces capability, a recent feature with v10. It is a way to organise related assets from multiple projects: 

Finally, other best practices such as naming conventions and documentation are also useful to get projects organised. Some suggestions on this page: https://knowledge.dataiku.com/latest/courses/governance/collaboration/best-practices-collaboration.h....

I hope this helps.