Logarithmic scale

[Mrinktvis]

Hello,

First of all thank you for this very useful tool to have easy insight in data.

Some data I use very often is normally displayed in a logarithmic trend instead of linear. I would like to see the ability to use a logarithmic scale.

With kind regards and thanks a lot!

[Waveworks]

Hello,

Thank you for the suggestion - that makes sense. I'm adding it to the list of feature requests.

For now you can use a formula like y_log = log(y) to visualize the logarithm of y. I understand it's not exactly what you want since it won't show the logarithmic scale and grid, but hopefully it helps a bit while waiting for this feature to be implemented.

Best regards,
Gustaf

[Peal]

Hello,

i've got a request that might have a similar solution / poses a similar problem for implementing:
i want to use the data from one column as the x-values for others.

Simplest example: coordinates (not my use-case)
To be able to load a file "circle.csv" with x- and y values, that can be set up in such a way that it draws a circle in the viewer.

[Waveworks]

Hi again Peal! Picking a variable to use as X is actually my most common request, and it's on the top of my list. Long story short, I need to rewrite quite a lot to make it work, so that's what's holding me back, but I have a plan for it and I hope I can release it during 2025.

However, your circle example (assuming you mean a csv file with points along the circle) probably won't be supported, as long as Flow only supports line plots. The circle would be more of a scatter plot - also very useful, but not in my current roadmap. For line plots, I think the X values will need to be monotonously increasing.

Adding yet another vote for using a variable for X!

Best regards,
Gustaf

[Peal]

I need to rewrite quite a lot to make it work

That's what i thought, the rendering currently seems pretty fixed (i can guess how you implemented the current X-axis scaling).
And it's the reason i'm telling you about this use case, so you can make the new implementation with this update in mind.

the X values will need to be monotonously increasing

According to my experience with sensor data, you can only hope for the time variable to be monotonously increasing.
With everything else (temperature, distance, brightness, ...) the sensor noise will more or less guarantee that you get some backwards jumps in the data (especially the raw data in a .csv).

This doesn't mean an implementation needing a monotonously increasing variable is useless, i just want to suggest to keep the following update in mind when implementing it

[Waveworks]

I see, yes I see them as two quite different features: Line plots with a time variable on the X axis, and scatter plots where you choose any two variables for the X and Y axis. I understand now that you're voting for scatter plots, possibly with the points connected by lines. It has been noted, and I'll at least take it into consideration now that I update the architecture.

Thanks again for the feedback!

Best regards,
Gustaf

[Peal]

scatter plots, possibly with the points connected by lines

Exactly, that's the goal.
(And written like this, it's really obvious that i revived the wrong forum topic, sorry)

plots with a time variable on the X axis

plots where you choose any two variables for the X and Y axis

It seems i'm really underestimating the complexity, because those two look pretty similar to me.
Especially because you need user input to define which data column should be time/x anyway.

I'll at least take it into consideration now that I update the architecture

Thank you and good luck with the update