Structured Streamlit Application Framework in Snowflake

Developing a Structured Streamlit Application Framework in Snowflake

Building scalable data applications can be challenging. At GitLab, our Data Science team recently took on the task of creating a structured framework leveraging Streamlit within Snowflake, designed to streamline development and improve user experience. Here’s how we approached it and what we learned along the way.

Why Streamlit and Snowflake?

Streamlit is known for its simplicity in building web apps for data science and machine learning projects. Snowflake, on the other hand, offers a powerful platform for cloud data management. Combining the two enables data teams to deliver robust applications directly inside the data warehouse, enhancing scalability, governance, and performance.

Designing with Structure and Modularity

We identified early that scalability and maintainability required a structured codebase. Our approach involved defining a centralised page router, consistent directory structure, and a reusable config system. This modular layout not only enables clearer code ownership but also simplifies debugging and collaboration.

Creating the Page Architecture

An essential component of the framework is the navigation system. We designed a page router to support multi-page applications, where each page acts as a self-contained module. Leveraging Streamlit’s native support for page navigation, we implemented a clean and efficient method to control the flow and UI of the apps.

Reusable Configurations and Assets

We defined shared resources such as CSS, logos, and configuration values in central repositories. Using a YAML-based configuration allows for easy overrides per environment, while consistent visual design enhances branding and usability across multiple applications.

Streamlining Deployment Within Snowflake

Snowflake’s support for Streamlit apps simplifies deployment. All required files and dependencies are stored directly in Snowflake, reducing external resource reliance and ensuring the application meets strict data governance requirements. This architecture also speeds up onboarding new apps or developers to the platform.

Lessons Learned

Implementing this framework not only improved our internal development process but also provided valuable lessons—primarily, the importance of early investment in structure and automation. Having a reusable application skeleton ensured that new apps could be created in hours, not days, and always aligned with internal best practices.

Whether you’re aiming to streamline internal reporting dashboards or productise your machine learning insights, combining Streamlit and Snowflake offers a powerful foundation.

Looking to build or scale your own data applications on GitLab and Snowflake? Our professional team at IDEA GitLab Solutions offers expert consulting, implementation, and licensing services across the UK, Czech Republic, Slovakia, and beyond. Let’s accelerate your transformation with robust DevSecOps solutions today.

• GitLab Supports Flexible Deployment as Atlassian Ends Data Center Offering
• GitLab Recognised as a Leader in the 2025 Gartner® Magic Quadrant™ for DevOps Platforms
• GitLab and Accenture Announce Global Reseller Partnership
• A Developer’s Guide to Building Secure Retail Applications with GitLab
• GitLab Docs Site Gets a Sleek New Design