Do you dashboard?

Hey there,

It’s Robin from CFD Engine & I’ve got a question for you this week…

“Do you have a CFD dashboard?”

By which I mean a central place where you display the key metrics for your simulations.

I’m asking because I don’t.

But, I’m seeing them pop up more frequently & wondered if the cool kids have them already 🤔

I’ve always said that most CFD engineers do essentially the same thing over & over (with some small changes – we’re not totally mad). And so I can see how it would be beneficial to build something that parsed your project data, extracted key metrics & displayed them in an intuitive (& possibly interactive) way.

I like it. I get it. But I’ve not done it. Have you? Let me know.

Background

This train of thought was prompted by the confluence of several LinkedIn comments & posts:

Max lamenting that our CFD data doesn’t get the eyeballs it deserves & therefore we’re missing out on diverse perspectives & ideas;
Simon sharing a dashboard he built to explore the outputs of a Bayesian optimisation, hosted (past tense) on AWS;
Astrid giving us a guided tour of her dashboard exploring data from a multi-objective optimisation;

I really like the idea of more project members playing with the CFD data & coming up with new “what ifs”.

I like that the users don’t need any special software installed on their machines.

I like that the data has been pared back to the key metrics – no one wants to dig through ALL of the data that their CFD run generated.

And I really like dashboards that can work on tablets, I can’t remember the last time I had a decent idea while sat at my desk.

I think I’m selling this idea to myself 😳

My own workflow

The closest I get to a dashboard is to strip key metrics (usually forces, flow rates & pressures) using command-line tools (grep, sed & awk anyone?) & present them nicely in a Google sheet.

It’s not exactly automatic – it’s a bit duct-taped together behind the scenes – but it works & ticks some of the boxes above.

My own next step would be to use the Google Sheets API to automatically add the data at the end of a run. Or maybe to see if my favourite friendly database, Airtable & its API, would be better suited for the job 🤔

Hmmm, perhaps I do need to dig deeper?

Shiny Objects

I’ve nerd sniped myself now – if I’ve also got you thinking, then here are a few cool tools to check out:

Streamlit - an open-source Python library for easily making data-driven web apps. Almost looks too good to be true (& I don’t even Python). Plus it’s THE nicest open-source-related website that I’ve ever seen 👀 This is the one that Astrid used.
Dash from Plotly - similar deal, open-source libraries to bridge the gap between your data & a web app. Also includes modules for manipulating VTK data with a couple of OpenFOAM examples tucked away too – looks nice. This is the one that Simon used.
D3 - an open-source Javascript library for turning data into web-ready visualisations which can be made interactive, zoom-able & more. Check out Observable while you’re at it for lots of lovely examples.
Datawrapper - this one’s focused on creating beautiful charts, tables & maps (the sort of stuff you might see in the Economist or NYT) & it gets expensive, fast - but oh so nice to look at. Check out their blog series on “Datavis Do’s & Don’ts” if you want to nerd out a little bit more.

There are LOADS of others, but that should be enough to keep us all occupied for a little while – sorry 🤷‍♂️

Over to you

While I’m off chasing new (to me) data visualisation libraries, let me know what you do with your data.

Have you built your own data dashboards? Did you go for a web app, or are you a spreadsheet ninja?

Keen to hear what you do & your experiences with tools like these. What’s good? What should I be checking out? Hit reply, I’d love to know more.

Until next week, stay safe,