pgVis - Simple Visualisations For PostgreSQL
I'm excited to announce the first version of pgVis a PostgreSQL extension for building simple visualisations dashboards with SQL.
pgVis aims to make it easy to express data visualisations directly from SQL queries. Letting you quickly visualise some data for adhoc reports in
psql or to share reporting dashboards in your organisation via
pgvis-server . Either way pgVis is designed to be PostgreSQL centric and to fit with your existing database workflows.
pgvis-server or statically saved via
psql or any other PostgreSQL client.
You can find out more on https://pgvis.org/ .
A Quick Demo
WITH write_4k AS (
-- The data analysis query we're going to visualise
SELECT device, buffered, direct
-- Define the visualisation
-- Add a bar chart
vis.bar_chart('SSD Write Throughput (KiB/s)',
-- Chart dataset, serialising as a JSON array
vis.dataset( (SELECT jsonb_agg(q.*) FROM write_4k q) ),
-- Add two data series onto the chart
vis.series('Buffered IO Throughput', 'device', 'buffered'),
vis.series('Direct Throughput', 'device', 'direct')
-- Hide the dashboard header
The idea behind pgVis came from some data analysis I was doing for a project and a number of situations I've run into in the past.
Working with PostgreSQL normally means you end up playing with data fairly often. In general PostgreSQL is my go to data analysis tool, I'm much more comfortable analysing and manipulating data in SQL than I am in say a spreadsheet. However so many times I find that I want to produce a simple chart or little report to share with people and find myself having to export data into another tool.
I've also worked in plenty of teams where people were comfortable building complex reports in SQL but we lacked a simple tool to display the information to other teams and management. Of course there are other tools for building dashboards and running BI, but often they've not fitted with our workflow.
So a simple tool which could be used for visualisations one off, shared or even embedding, could be of use to other people too.
This release is very much minimum viable, it demonstrates the concept and ethos of how I want this tool to work and focus upon, while providing enough features to be useful to me and others. This initial release focuses on getting things working end to end and simplicity. The currently supported chart visualisations are simple, but there is a lot more power under the hood by constructing the right options. Likewise there has been very limited focus on supporting dynamic and filterable visualisations at the moment.
There are certainly more features I've got planned and a range of ideas in mind. In the next release I'll be aiming to have more support for dynamically updating and filtering data via the pgVis server.
I'd also be very interested to hear peoples thoughts, especially features that people could see being of use to them.
Feel free to message me on Mastodon: @firstname.lastname@example.org
I'm planning on accepting feature requests and issues on Gitlab.