You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am thankful for the opportunity to share our feedback as part of the final review (#476) and I appreciate the effort DIME is putting in disseminating these valuable guidelines and resources.
The chapter addresses a crucial part of a project's success: collaboration. Here are some ideas, especially coming from the angle of the Data Partnership. I'd be more than happy to collaborate.
Ideas
More often than not, relying on absolute paths causes trouble. It will almost guarantee your code won't run on computer other than yours.
Ensure you know how to get the \textbf{absolute file path} for any given file.\index{absolute file paths}
It is recommended to check the Bank's stance on Dropbox. Alternatively, the Bank supports OneDrive with the advantage, other than being official and offering up to 5TB per account, of ensuring data classification (Official Only, Confidential, Strictly Confidential).
It would be beneficial to have additional step-by-step examples on how to set up the many recommendations on the chapter. More can be found at DIME Wiki, but the intended audience might find helpful to have quick guides or more references to tutorials.
The book touches on a super important point when it comes to team communication and decisions. However, the section might need elaboration. Using tools like GitHub or Dropbox won't help much unless the team adopts an effective approach to project management. For example Agile, Agile-like, Scrum, Kanban. Of course, GitHub does support amazing features like GitHub Projects that can dramatically improve the team's performance (and sanity). In a nutshell, what's important here is not the tool, it is the process.
Good collaboration tools are workflow-oriented systems
that allow the team to create and assign tasks,
carry out discussions related to single tasks,
track task progress across time, and quickly see the overall project status.\index{task management}
They are web-based so that everyone on your team can access them simultaneously
and have ongoing discussions about tasks and processes.
Such systems link communications to specific tasks so that
related decisions are permanently recorded
and easy to find in the future when questions about that task come up.
Choosing the right tool for your team's needs is essential to designing an effective workflow.
What is important is that your team chooses a system and commits to using it,
so that decisions, discussions, and tasks are easily reviewable long after they are completed.
Some popular and free collaboration tools that meet these criteria are
GitHub and Dropbox Paper.
Any specific list of software will quickly be outdated;
we mention these as examples that have worked for our team.
Different collaboration tools can be used different types of tasks.
Our team, for example, uses GitHub for code-related tasks,
and Dropbox Paper for more managerial tasks.
GitHub creates incentives for writing down why changes were made
in response to specific discussions
as they are completed, creating naturally documented code.
It is useful also because tasks in GitHub Issues can clearly be tied to file versions.
On the other hand, Dropbox Paper provides a clean interface with task notifications,
assignments, and deadlines,
and is very intuitive for people with non-technical backgrounds.
Therefore, it is a useful tool for managing non-code-related tasks.
Probably out of scope, but it would be great to have a section on cloud computational environments and resources, such as JupyterHub, AWS Sagemaker or Google Colab.
Probably out of scope, but Python is a dispensable part of a modern analytics stack and there are considerations that might be useful when using Python or, more specifically, working on a data science project.
Probably out of scope, same goes for containerization with Docker.
The text was updated successfully, but these errors were encountered:
I am thankful for the opportunity to share our feedback as part of the final review (#476) and I appreciate the effort DIME is putting in disseminating these valuable guidelines and resources.
The chapter addresses a crucial part of a project's success: collaboration. Here are some ideas, especially coming from the angle of the Data Partnership. I'd be more than happy to collaborate.
Ideas
More often than not, relying on
absolute paths
causes trouble. It will almost guarantee your code won't run on computer other than yours.dime-data-handbook/chapters/2-collaboration.tex
Line 79 in ba0105d
It is recommended to check the Bank's stance on Dropbox. Alternatively, the Bank supports OneDrive with the advantage, other than being official and offering up to 5TB per account, of ensuring data classification (Official Only, Confidential, Strictly Confidential).
dime-data-handbook/chapters/2-collaboration.tex
Lines 98 to 101 in ba0105d
It would be beneficial to have additional step-by-step examples on how to set up the many recommendations on the chapter. More can be found at DIME Wiki, but the intended audience might find helpful to have quick guides or more references to tutorials.
The book touches on a super important point when it comes to team communication and decisions. However, the section might need elaboration. Using tools like GitHub or Dropbox won't help much unless the team adopts an effective approach to project management. For example Agile, Agile-like, Scrum, Kanban. Of course, GitHub does support amazing features like GitHub Projects that can dramatically improve the team's performance (and sanity). In a nutshell, what's important here is not the tool, it is the process.
dime-data-handbook/chapters/2-collaboration.tex
Lines 140 to 187 in ba0105d
The text was updated successfully, but these errors were encountered: