Gustaf Alströmer

Evidence - Beautiful reports with just SQL and markdown

by

Evidence is an open source, code-based business intelligence tool. It enables analysts to build automated reports by writing SQL and markdown, instead of using a drag and drop interface, like you’d find in a traditional BI tool.

Add a comment

Replies

Best
Rupjyoti Nath
Liked the concept of using markdown. Any chance of supporting MongoDB ? At least a workaround to work with MongoDB would be helpful. MongoDB charts are fine but quick reports in markdown with descriptions would be awesome!
Sean Hughes
@rupjyoti_nath it’s not something we’ve thought about yet - we’ll need to investigate. We’ve opened an issue on GitHub here if you want to track updates on this: https://github.com/evidence-dev/...
Brice Maurin
Seems great ! Did you have a look at the filters and advanced syntax you can use in streamlit?
Adam McAskill
@b_maurin thanks Brice! Are you referring to things like sliders and multi-selects? We haven’t released any interactive components like that, but they are on the roadmap.
Brice Maurin
Yes ! The filters, sliders, option boxes, select, forms...everything that could transform static reports into interactive ones !
Adam McAskill
@b_maurin 100% we’re going to add those types of things to our component library and make it really easy. If you know a little HTML and JavaScript you can add that type of interactivity today. Writing HTML input tags directly into your markdown just works. I’ll spin up an example for you. 😎
Adam McAskill
Hi everyone, I’m Adam -- one of the makers of Evidence, along with my partner Sean (@hughess). We’ll be hanging out in the comments to answer any questions. Sean and I are data analysts at heart. We started our careers doing analysis in excel and R, and writing up our work in word documents and powerpoint decks. Eventually we needed to graduate to something more scalable, and we moved on to using the tools that would later become known as “the modern data stack”. Tools like BigQuery, dbt, Looker, and Mode. We gained a lot from moving to the modern data stack, but we also felt like we lost something... Dashboards, it turns out, are poor substitutes for handwritten analysis. A handwritten report by an expert analyst presents its most important conclusions, then takes its audience through a persuasive line of reasoning to support them. It drives its audience toward a set of next steps. It’s filled with supporting information, context, definitions, and direction on where to learn more. It builds understanding, trust and confidence. It's just too hard to consistently deliver that type of experience using a drag-and-drop BI tool. Producing content is too slow, maintaining it is too unwieldy, and the "charts on a grid" form factor is too inflexible. Best case, you put some filters on your dashboard and ask your audience to do the analysis themselves. Evidence takes a very different approach. By replacing the drag-and-drop workflow with a flexible markup language, Evidence enables analysts to deliver automated reports that feel like they were handwritten. Want to generate sections in your report for each region of your business? Use a loop. Want to only show a graph when last month's margins were off target? Write a conditional. Need to add commentary that only appears when someone is looking at January's results? No problem. Evidence combines the best features of a static site generator like Jekyll, and a notebook like Jupyter into a package that’s easy for SQL analysts to learn. Sean and I have benefited tremendously from the open source community. Incredible tools like R Studio and dbt have had a huge impact on our careers and our lives. We’re making Evidence open source to carry on that tradition. We aim to build a tool that will appeal to the most sophisticated data teams in the world, and will also feel approachable to someone who is writing her first line of SQL. We have a long way to go but we would love to have you be a part of this journey.
Dmitry Goloktionov
That's a great idea. This framework will simplify the creation of reports.
Sean Hughes
@dmitry_goloktionov thanks - let us know if you have any questions!
JJ Tang
such a good idea, I wish I came up with it
Sean Hughes
thanks @jj_tang !
Lior Chen
Wow Great Idea, loved it!
Sean Hughes
@lior_chen thanks!
Harry Brundage
This is so awesome! Seems like the perfect complement to dbt -- a developer native way to build reports with full control. What's next on the roadmap for you folks?
Adam McAskill
@harrybrundage 100% we think teams who are using dbt will really like how evidence works, and we plan on building some slick integrations with dbt. Our immediate roadmap is all about tightening up the core open source experience and getting it out of alpha, and adding support for more databases.
Nate Matherson
This is super cool!
Daniel James
Congrats on the launch !
Olamide Akomolafe
This is an amazingly simple idea and product. Absolutely love...
Sean Hughes
thanks @theakomolafe - glad you like it!
Udhaya Kumar
Simply wonderful!
Kelly Suzan
Interesting!
Sean Hughes
@kellysuzan let us know if you have any questions!
Betsey Millership
Indeed great!