Notes on the Perfidy of Dashboards

From TFA... >every dashboard is a sunk cost

>every dashboard is an answer to some long-forgotten question

>every dashboard is an invitation to pattern-match the past

>instead of interrogate the present

>every dashboard gives the illusion of correlation

>every dashboard dampens your thinking

I disagree with this on all counts. A dashboard is a way to view multiple disparate metrics in a single place. Whether they are correlated isn't important(but it is helpful).

And the author doesn't stop there...

> They tend to have percentiles like 95th, 99th, 99.9th, 99.99th, etc. Which can cover over a multitude of sins. You really want a tool that allows you to see MAX and MIN, and heatmap distributions.

They "tend to"? "You really want"? The author is confusing their own failures/gripes around the concept of dashboards with the world's experience with dashboards. By the end of the article, I was shocked they weren't selling something.

I agree with both viewpoints in certain ways, a lot of times visualization surfaces issues that would otherwise go unnoticed but I think the important point is that being able to drill down is important for context.

The tld is apt, wtf?

Dashboards are invaluable. Humans can intake a lot of data from images and there is not better way to grok data than a graph.

We've spent a lot of time building Grafana dashboards and they've been extremely helpful with debugging. It doesn't solve all problems but it certainly helps narrow down where to look.

Sure, we still look at log files, use htop and a lot of other tools, but our first stop is always Grafana.

I suggest the almost any book by Edward Tufte. There you'll see the beauty and value of visual information.

The main problem is that it's not easy to drill down into a metric in most BI tools because the connection between the dashboard and the source data is usually missing. Looker is one of the first companies that target this specific issue; you transform, model the data, define your metrics before creating your first dashboard. It takes too much effort (also money) to create a (basic?) dashboard because it's not just a "dashboard".

Instead, as data analysts, we usually want to write a bunch of SQL queries, create charts from them and expose the data to our business stakeholders. While they can see the underlying SQL queries of the metrics, it's not easy for them to modify these SQL queries, so they often get lost.

The dashboards have a long tail. For me, you need to get these four steps done beforehand:

1. Move all the company data into a data warehouse and use it as the single source of truth.

2. Model the data with a transformation tool such as dbt and Airflow.

3. Define metrics in one place on top of your data models in a collaborative way. (This layer is new, and we're tapping it at https://metriql.com)

4. Use an interactive BI tool that lets you create dashboards on top of these metrics with drill-down capability.