Project templates best practices
Learn how to utilize project templates to jumpstart your data science initiatives.
Templates are a great way to start your project. The default project template provides a project based on the CRISP-DM methodology, certified by Data Science Process Alliance, commonly used in the data science process. However, your organization may have a unique internal process. You can utilize the following best practices to maximize the benefits of using project templates.
#1 Customize the default project template for specific needs
Every organization is unique and may require different processes. For this reason, we encourage you to customize your template to fit your organization's unique needs. For example, you can create project templates for Analytics, Operations, Machine Learning, and more.
You can customize your project template by adding or removing phases, requirements, and documentation as needed. View the Setting up a Project guide for more information on completing these actions.
Once you have a customized project template that works for your organization, we recommend you Duplicate this template for future projects.
#2 Color code text placeholders in templates to display information
To make your project template easy to follow, you can color code the text within your template to highlight the phase's descriptions, instructions, or purpose and use another color to represent different information.
For example, each phase within the default project template has a description and purpose written in blue that you can edit or delete. Anything written in black is used as phase documentation outlines.
#3 Duplicate your projects
After creating a project template that can work across multiple projects, you can Duplicate the project. This will create a new project that preserves the project template's phases, requirements, and documentation. This will simplify the process as you will not need to rebuild projects from scratch. You can duplicate a project by selecting your Project Settings > Duplicate Project.
When duplicating a project, the new project will not preserve the iterations, models, and datasets.
#4 Use common workspaces to share project templates
Share your customized project templates with organization members or teams that could use them within their workspaces. To share project templates, create or use a Workspace that includes all members with who you want to share the template(s). The members can then duplicate the template and select the workspace where they want to place the project.
#5 Pre-define phase requirements within your template
You can pre-define the requirements needed to complete the phases within your project template. Once you have shared and a user received this template, they would have the requirements already included to complete the project phase.
For example, the data preparation phase in the default project template, is pre-defined with the requirements to complete that phase.
You can enable the "Set requirements as sections for iterations" toggle in which the requirements would be used as section names and descriptions for future iterations of the phase.
#6 Move project templates to designated workspaces
When creating a project template for a specific group, you can move the customized template from its current workspace to the desired workspace. Moving a project allows you to keep all project assets and information, including datasets, models, code, and iterations. This allows you to pre-upload datasets and other assets within your template before sharing.
You can only move projects to workspaces that you are a member of.