Building AI Apps with H2O Wave by Martin Turoci
Overview
H2O software engineer Martin Turoci gives a live coding demo of H2O Wave, a Python framework for building web-based AI apps without writing HTML, CSS, or JavaScript. Starting from an empty project, he builds a working app with buttons and text input, shows how to browse Wave's example gallery, and demonstrates theming and deployment options. He also shares that H2O's own Hydrogen Torch UI was built by data scientists with no web development background in two to three weeks.
Key takeaways
- How to install H2O Wave and scaffold a first app using the pages-and-cards model, all in Python
- How to handle interactivity: capturing button clicks and text box input through Wave's query object
- Where to find Wave's built-in widgets, native visualizations, and support for Altair, D3, Matplotlib, Plotly, and Vega
- How to browse and run the 200-plus example apps bundled with Wave via the wavefetch command
- How to theme an app with built-in color schemes or a custom style sheet, and integrate external JavaScript if needed
- Deployment options for a finished Wave app, including Heroku, AWS EC2, and H2O's own cloud environment
Start owning your AI