Streamlit: The Best Developer Experience for Data Apps
Today we are excited to announce that Sequoia is leading Streamlit’s Series B, alongside existing investors GGV and Gradient Ventures. Streamlit gives data scientists superpowers: with just a dozen lines of python, any data scientist can turn a local project into a beautiful, interactive data application, no front-end code required.
We have been tracking Streamlit since the open source project launched in 2019. Out of the gate, Streamlit was popular with data scientists, who used Streamlit for creating all sorts of data apps from stock portfolio composition analysis to fake horror movie plot generators to COVID data visualizations to GPT demos.
In 2020, several months after the open source release, we noticed an exciting new development. Data scientists were not only tinkering with Streamlit for their personal side projects, but they were starting to use the tool to build robust Streamlit applications at work, such as retail basket analysis, call center AI, marketplace operations optimization, and supply/price forecasting. Some of these data applications were suddenly being used on a daily basis by hundreds or thousands of business users.
Why We’re Partnering
Data teams have more insights than ever before. And business users are ravenous to consume it. Unfortunately, the gap between the two is wide, and the information that our data scientists hold at their fingertips far exceeds what is readily accessible to business users. The interface between data teams and business users remains an unsolved problem.
Over the last two decades, the field of Business Intelligence has attempted to democratize data through static visualizations on top of relational databases. Companies like Tableau, Looker and Microsoft PowerBI have made it easy enough for non-technical users to make some sense of data through rich visualizations and familiar user interfaces.
We are in new and exciting times for data science, compared to the BI tools of the 2000s. Data volumes are compounding and increasingly unstructured; models have progressed in sophistication to complex neural nets and machine learning; and code — crucially, python — has become data scientists’ preferred medium.
We believe these shifts require a new way of interacting with business data. Static visualizations on relational databases (a.k.a. the current BI tools) are no longer sufficient to express all of the insights and predictions that we can glean through data. Rich, interactive data applications are the answer. We believe that the best developer experiences (fast; intuitive; python-native) will win and that crafting a beautiful application experience for end users is non negotiable.
Enter Streamlit. Streamlit has taken a developer-first, python-native approach to BI. By focusing on the developer experience and making it simple, even magical, to launch a rich data application into production, Streamlit is positioned to become the favored tool of data scientists and the new interface between data teams and business users.
What’s Next for Streamlit
Over the last six months, the Streamlit team has been hard at work developing and beta testing a commercial product: Streamlit for Teams. Streamlit for Teams is in closed beta and will launch later this year.
While any data scientist can build a Streamlit app using the open source project, deploying and sharing these applications is hard. It requires provisioning infrastructure, managing CI/CD, hosting the application, and gating and managing application access. So far, users have jumped through hoops to deploy Streamlit applications into production, by hacking together features like password authentication.
Streamlit for Teams takes all the goodness from the open source and makes it available as a fully managed service. It manages hosting, deployment, collaboration/sharing, and data security, with more advanced enterprise features available including SSO and data governance to manage application access.
We believe Streamlit for Teams will reduce the barrier to deploying data applications inside an organization and give data scientists the superpower of application development.
Streamlit is ushering in a new category of data-rich applications, from personal passion projects to critical business functions. We believe the company has a chance to scale up insights from a single data scientist to entire business teams and shake up the $25B business intelligence market. We are thrilled to join Adrien, Amanda, Thiago and the entire Streamlit team on this journey.
If you’re interested in joining the Streamlit team, they are hiring across Engineering, Product, Design, Marketing, Sales and Customer Success. Data scientists interested in trying Streamlit can get started from the open source or request access to Streamlit for Teams.
