Concept | Webapps#
Webapps are custom applications hosted by Dataiku, which allow you to extend Dataiku’s visualization capabilities. While Dataiku offers many native charts, webapps provide a way to build custom interactive experiences for your end users.
Why use a webapp#
Webapps bridge the gap between complex data science and business value. While Dataiku offers extensive native visualization tools, a webapp might be the right choice depending on your specific requirements and coding expertise.
Because webapps are code-based, a key factor in choosing this route is the creator’s proficiency in the relevant language (such as Python, R, or JavaScript).
A webapp could be a suitable option if your project requires:
Requirement |
Description |
|---|---|
Custom interactivity |
Dashboards that update based on complex user inputs or logic not found in native charts. |
Specialized visualizations |
Use of specific libraries (such as Leaflet or Plotly) for unique data storytelling. |
Operational frontends |
Custom interfaces that allow users to trigger backend processes or interact with data. |
Webapp types#
Dataiku supports several frameworks, allowing you to choose the one that fits your coding expertise.
Code-first frameworks (Python and R)#
These frameworks are ideal for data scientists who want to build apps using only their preferred language.
Dash is a productive Python framework for building analytical web applications.
Bokeh is a Python library for creating interactive visualizations for modern browsers.
Shiny is the standard for building interactive web applications using R.
Streamlit is a popular Python framework for turning data scripts into shareable webapps in minutes.
Standard webapps (HTML/JS)#
For full control over the user interface, you can build webapps using HTML, CSS, and JavaScript. These apps can also include a Python backend (using Flask) to handle data processing and API interactions.
See also
To learn how to build a standard webapp, see HTML/CSS/JS: your first webapp in the Developer Guide.
Visual webapps#
These are code-based webapps (developed as plugins) that offer a visual interface for non-coders to select datasets and parameters without writing code.
Create a webapp in Dataiku#
You can access webapps from the top navigation menu of a project, by hovering over the Code menu and clicking Webapps.
To develop a webapp, you can start from scratch or from a template. Dataiku takes care of the hosting, making the app available to anyone with access to your project.
Webapps vs. Dataiku apps#
It’s important not to confuse webapps with Dataiku apps. These are two different tools.
Webapps |
Dataiku apps |
|
|---|---|---|
Development |
Code (Python, R, JS) |
Visual UI (no code) |
Primary goal |
Custom UI/Visualization |
Reusable project automation |
User experience |
Free-form app |
Form-based automation |
See also
Learn more about the difference between Dataiku apps and webapps in Concept | Dataiku apps.
Share and publish a webapp#
Before others can interact with your webapp, you must manage its accessibility and visibility. In the Dataiku UI, sharing and publishing are distinct actions with different purposes.
Share a webapp#
Sharing makes a webapp available to other projects.
To other projects: Click Share to make the webapp available for use on dashboards and insights within other projects.
User permissions: To control which specific users or groups can see or edit the webapp, use the Go to Project security settings link.
Publish a webapp#
Publishing makes the webapp visible as a consumable insight.
Dashboards: Select Publish > Dashboard to display the webapp as a tile on a dashboard.
Workspaces: Select Publish > Workspace to add the webapp to a dedicated workspace for business users.
Public URL: For broader reach, administrators can make webapps public, allowing access via a standalone URL outside of the Dataiku interface. For more information, see Public webapps in the reference documentation.
Next steps#
You have learned about webapps and how they work in Dataiku.