Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Launch HN: Evidence (YC S21) – Web framework for data analysts

news.ycombinator.com

Post reply on HN
Print view View original

11–20 of 94 posts

Re: Launch HN: Evidence (YC S21) – Web framework for data analysts

#12

Post by nonameiguess »

Have you heard of knitr (https://yihui.org/knitr/)? It's the gold standard as far as I'm concerned for dynamic report generation that needs to run code. Since it supports running arbitrary shell commands, it can already be used to query remote databases as long as you have a CLI to query them with. Combined with RMarkdown (https://rmarkdown.rstudio.com/), which augments Markdown with support for LaTeX typesetting, it's the ultimate toolset for doing this kind of thing. You can read a blog post here on how to use knitr within RMarkdown: https://kbroman.org/knitr_knutshell/pages/Rmarkdown.html

I'm not trying to be a downer, but it seems like your product is just duplicating the functionality of these existing products but does less since it only supports SQL and Markdown.

I guess you autogenerate charts, but it says you're targeting a technical audience that is presumably comfortable calling functions in Python and R for graphical data visualization.

This is nitpicky, and I'm sure you have some command line option to choose another port (though your "get started" doesn't show how), but mdbook also uses 3000. I'm sure they probably weren't the first to default to that, either.

I hope this doesn't come across as downplaying your product. It looks nice. I just don't see what you offer here that can't already be done with existing data ecosystem tools. I was using RMarkdown with knitr to generate all of my papers when I was an ML grad student years ago. It felt back then like I was the only person at Georgia Tech who realized these tools existed, and now it still feels that way.

Re: Launch HN: Evidence (YC S21) – Web framework for data analysts

#14

Post by edusoftwerks »

nonameiguess wrote: post #12

Have you heard of knitr ( https://yihui.org/knitr/ )? It's the gold standard as far as I'm concerned for dynamic report generation that needs to run code. Since it supports running arbitrary shell commands, it can already be used to query remote databases as long as you have a CLI to query them with. Combined with RMarkdown ( https://rmarkdown.rstudio.com/ ), which augments Markdown with support for LaTeX typesetting…

The problem with these tools is they only work sometimes, and when they do, its because you spent a whole day configuring your environment.

Re: Launch HN: Evidence (YC S21) – Web framework for data analysts

#16

Post by sails »

> In our experience, the BI layer is the weakest part of the modern data stack. The BI layer has a poor developer experience, and decision makers don’t really like the outputs they get

Totally agree, very interested in trying this out. FWIW I've tried and been frustrated by Looker, Metabase, PowerBI, Superset, Redash.

I do think that while dbt does a great job with dimensional modelling, the BI layer is still required to provide some aspects of metric modelling. Is this something that Evidence is looking to solve? From what I've seen it looks more to be a pure frontend visualisation rather than a tool for managing business metrics. Looker and Metabase do some good work in this metric management space, Superset and PowerBI much less so.

Re: Launch HN: Evidence (YC S21) – Web framework for data analysts

#19

Post by amcaskill »

101008 wrote: post #8

How do you handle live data vs fixed data? If I am making a report, I want the charts to remain static - if not, over time, they may not match with what is said on the report. Is there an option to, after saving the report or run the query, to make the values static forever?

This is an excellent question, and it's one of the areas where we think we can do some pretty novel things with our approach.

There are a two main cases of this idea that we have spent time thinking about.

1. Truly static report.

Here, you would need to condition your SQL queries so that they continue to return the same results over time. E.g. your `where` clause restricts the results to 'on or before' the day of writing. Evidence will continue to build the report on a schedule, but the results will never change so long as your historical data is constant. You can do this today.

In a future state, we've talked about rendering a snapshot of the report and checking that into version control, so that even if your underlying data is a moving target, you can hold onto what the report looked like at a moment in time.

We're kind of mixed on that idea of snapshotting reports themselves though, it's just so much better to build your data warehouse such that it is actually retaining the historical data, but we recognize sometimes that's not practical. TBD on that functionality.

2. Recurring report with static commentary

Here, you have a recurring time-bounded report, and you occasionally want to mix-in commentary that's only relevant for specific time periods.

With Evidence, (this part is from svelte kit) you can mix paramaterized pages, and static pages on the same route. So if you had 'monthly mrr growth report', you could use a parameterized page to generate the report for every month into history and into the future, AND, you could include versions of the report with hand-written commentary for any specific months where it was needed. So if someone navigates to the February 2021 page, they get the standard parameterized version, but if they go to January 2021, they get the handwritten January version that explains that there was an acquisition which drove the big pick-up in MRR.

This one is a bit tricky to explain, but we will build some examples.

Re: Launch HN: Evidence (YC S21) – Web framework for data analysts

#20

Post by joefigura »

Very cool! I struggled with this exact problem at a previous role to generate reports from the results of a simulation codebase. Ended up learning and using Django, which always felt overpowered for a simple static report. Excited to see how the product develops!
Post reply on HN