Evidence – Web framework for data analysts
When everything is under version control, you are less likely to ship an incorrect report
I discover a new startup every week and share it with you. This week’s featured startup is Evidence. Subscribe to get each and every issue.
What it does: Evidence is a static site generator for data analysts. This allows tech-savvy data analysts - who can also code - to create nice static reports with SQL + Markdown syntax with powerful templating tools and bring it all under versioning control.
Why it’s different: Most BI tools use a no-code drag-and-drop interface which does not allow for well-known software development practices such as versioning control, testing, and abstraction. BI tools are built that way because they assume that data analysts are non-technical. In Evidence’s experience, that assumption is no longer correct. Evidence allows you to create static reports that can be version controlled, tested and abstracted.
years building the data science group at a private equity firm in Canada. They started their careers doing analysis in excel and R, and writing up their work in word documents and powerpoint decks. After moving to a more modern and more scalable data stack, they felt that they had lost something along the way and realized that dashboards are poor substitutes for handwritten analysis.
Business model: The core framework will always be open source and eventually they plan to launch a paid cloud version of their product with enterprise features.
How it works: With Evidence, you can write SQL and MarkDown to create static data reports. It allows you to query your data from Markdown (with support for BigQuery, Snowflake and PostgreSQL) and include auto-generated charts and graphs.
Why it’s interesting: Evidence solves the kind of problem that you typically only experience by doing the actual work of a data analyst or working with them directly. Having worked intensively with data analyst teams in the past, I immediately recognized the pain points that Evidence was solving. I believe it should be a no-brainer for data teams who are eager to keep improving their productivity and reliability. There certainly is no shortage of BI tools, but very few actually focus on making the data engineering process itself better and up to par with well-established software development practices.