Page 1 of 1

Alternate wind graph thought.

Posted: Tue 31 Jan 2017 8:40 pm
by Phil23
Just been playing with the wind graphs in jpchart, (& probably out of my depth).

Wondering what thoughts are on this idea, if I can explain it so it makes sense.

A Scatter graph of average wind direction, starting out with say 3x3 plot points.

For Wind speed, could that value take control of the plot point size, so the vertical size of the plot point is controlled by wind speed.

So say plot point sizes might be:

0km/h = 3x3
10km/h = 3x10
30km/h = 3x30

etc...

There would need to be unused area above & below the graph, so the bars didn't get truncated, something like a -40 to 400 direction axis.

Does that all make sense, if it was doable, feasible, I'm assuming it would look something like this, without the shading & symmetrical.
Y-Axis displacement indicating wind direction of course.
Capture.JPG
Does anyone know a correct term for this type of graph?
Think it would make a superb visualisation of wind speed & direction.

Phil

Re: Alternate wind graph thought.

Posted: Tue 31 Jan 2017 9:49 pm
by mcrossley
I'm not quite sure I understand you explanation, nor what the graph is showing :lol:

But you have three axes of data, so either a 3D plot is required, or a 2D with colour say representing the third data 'direction'.

Remember that the point of a graphic is that it should be clear what it is representing and intuitive to 'decode' - though there are exceptions, where technical data can have a more complex representations that have to be learnt.

Re: Alternate wind graph thought.

Posted: Wed 01 Feb 2017 12:41 am
by Phil23
Didn't thik I was explaining it that well.

Only a 2D plot, but more like a stacked area plot, with speed stacked on the top of area.

I've tried to creat an Example in Excel with some real data.

For the sake of scaling, I multiplied wind speed by 5, then stacked it on direction.
In reality it would need an offset for direction to be at the centre of the Red plot.

The thickness of the red plot would indicate the speed, it's Y-Axis position the direction.
Smooth curve would be better, but can't seem to do that with Stacked Area in Excel.

Does this help explain it better?

Obviously it would need some sort of scale for speed, & auto scaling the speed stacking to suit.

And replacing 0-360 with N-S-E-W etc.

Phil.

Edit:- The Example below is a 48 hour plot.
Capture.JPG

Re: Alternate wind graph thought.

Posted: Wed 01 Feb 2017 6:43 am
by Phil23
Here's another example.

This time I used a VLOOKUP on the wind speed to bet the "Bar Height", IE the second stacked value.

The first Stacked area is Direction-(BarHeight/2).

The second stacked area is the wind speed.

The first area has no fill in Excel.

Normally in Excel it would look like the 2nd image if both Series had a colour fill.

With the Axis going from -45 to +405 I could have a maximum "Bar Height" of 90


Phil.
Capture1.JPG
Capture2.JPG

Re: Alternate wind graph thought.

Posted: Wed 01 Feb 2017 7:57 pm
by Phil23
Maybe a more visually appealing option is to have 2 axe's, Speed & direction.

A smooth curve for direction & an expanding area in the centre for speed.

Something like this, with direction added...
And a Speed axis on the left, direction on the right...
Capture.JPG

Re: Alternate wind graph thought.

Posted: Wed 01 Feb 2017 8:59 pm
by Phil23
Bit like this.

Would be good if something could ve done about the swings from say 359° to 1°, with say -22 to +22 for a slowly changing trend, to stop the direction curve swinging across the graph in those instances of small change.

Maybe based on running average.

Rambling out loud.....
Capture.JPG

Re: Alternate wind graph thought.

Posted: Wed 01 Feb 2017 10:13 pm
by mcrossley
That's way you have to use a scatter plot, otherwise graph plot will have a line from 1-360 and vice versa.
Or you have to detect the transitions yourself and break the data into sections so the plot restarts at each transition.