Embedded Analytics with Streamlit
Creation of native dashboards in a SaaS platform, overcoming Looker Studio's ACL limitations.
Context
Implementation of a fully customized and embedded analytics solution for clients of a SaaS platform, using Streamlit to bypass the permission limitations of Looker Studio.
Problem
The SaaS platform needed to offer dashboards to its clients, but Looker Studio only allowed access to users with emails from the same domain.
It was unfeasible to create and manage credentials for each client in the Google ecosystem.
The solution had to be customizable per client and integrated into the product experience.
Solution and Contribution
- I researched and validated Streamlit as the ideal solution to the problem.
- I developed the architecture to serve the Streamlit dashboards securely and in an integrated manner.
- I created the dashboards, using Python for data processing and visualization, with interactive filters and charts.
Diagrama de alto nível da arquitetura da solução do Embedded Analytics with Streamlit
Results and Impacts
- 100% customizable and integrated dashboards for SaaS clients.
- Elimination of Looker Studio's access and credential management barriers.
- Creation of a foundation for future data and data science products within the platform.
Technologies Used
Technical Details
Flexibility with Python
By moving away from Looker Studio, we gained the power of Python for data preprocessing. This opened up a range of opportunities to apply more complex transformations and, in the future, integrate machine learning models directly into the dashboards, creating a true data product.