Can a Dash dashboard be dynamically 'generated' with linked brushing/cross-filtering?

Dash Python
1

Hi!
I’ve only quickly went through the tutorials for Dash so I need some confirmation before I’m ready to commit on Dash to fulfill a goal. So basically I worked on a tool which drew specific common graphs using Plotly Python. So now I have to work on a dashboard feature where I can display all those graphs together with cross-filtering/linked brushing i.e. basically selecting a point in one graph would also highlight the same data point in other visualizations if possible. Also, there’s only one dataframe so that makes it simpler too.

Now my question is simply this; is it possible to achieve this with Dash Python?

In the tutorials, Dash seemed more for like static Dashboards where you already know how you want to layout your dashboard. Is there some trick I could use to achieve this? Because I only have to support like 15 visualization types so some kind of ‘hacks’ would be welcome too. I do not have to do anything detailed like saving/loading old graphs. It would be more like I have this list:
"line_chart: col1, col2
bar_chart: col2, col3
scatterplot: col1, col2, col3
"

and I would make new charts in Dash from scratch WITH linked brushing/cross-filtering or whatever it’s called, but basically just ;select in one, reflect in others’. If so, any helpful links would be appreciated as well, or a just a general direction on how I can achieve this.

Thanks in advance!

1 Like
2

Bump! (Hoping this is allowed). Still looking for an answer to this :frowning:

3

I think it should be possible. But it might require some hacking to avoid circular references.

4

While this topic is old, think I should just put this out here for any future Googlers. Yes, this is possible thanks to pattern-matching callbacks. I think I did an alright implementation of it.