Is it possible to generate chart with broken Y axis

Hi

I have to say sorry about my poor English, and here's my question, is it possible to generate chart with broken Y axis, just like the attached image, thanks for your reply.

chart with broken Y axis

One way to do it would be to use two left axes and draw only parts of them:

make /T City={"London","Paris","Madrid","Berlin"}
Make val={{3,5,67,4},{3,5,2,4},{3,5,3,65},{63,4,1,66}}
display val[0][],val[1][],val[2][],val[3][] vs city
appendtograph /L=left2 val[0][],val[1][],val[2][],val[3][] vs city
ModifyGraph axisEnab(left)={0,0.2}
SetAxis left 0,20
SetAxis left2 50,70
ModifyGraph axisEnab(left)={0,0.45}, axisEnab(left2)={0.55,1}, freePos(left2)=0

If you don't insist on the gap, a user tick wave might also do the job (but it somehow kills the message from the graph)

HJ

In reply to by HJDrescher

Another approach is to use the built-in WaveMetrics package for splitting an axis. In your Procedure file add

 #include <Split Axis>

Then look under the Graph menu for "Split Axis Control". Once the axis has been split you can alter the different split region ranges by selecting each region and using the standard GUI dialog. This also provides the (optional) diagonal split marks.

My question is more to answer "Why This Way?" than "How to Get it This Way?". Figure 3.15 of this book does a good job to illustrate why split axes graphs are misleading.

In that regards, have you considered simply using a log axis?

In reply to by HJDrescher

Thanks for your help.

And the reason for why this graph, our data are going to provide to some people who don't have  professional knowledge, so we decide to show original data that they could easily know what happening by our chat, I'll check out the book, maybe we can find out some better way to show our data, thank you.

In reply to by jjweimer

jjweimer wrote:
In that regards, have you considered simply using a log axis?

For the uninitiated, a log axis can be just as misleading, making large values appear to be similar to order-of-magnitude smaller values. Dealing with a great disparity in values in a transparent and easily-understood way is a very tough problem.

In reply to by johnweeks

johnweeks wrote:

For the uninitiated, a log axis can be just as misleading, making large values appear to be similar to order-of-magnitude smaller values. Dealing with a great disparity in values in a transparent and easily-understood way is a very tough problem.

Yes indeed.

Forum

Support

Gallery

Igor Pro 9

Learn More

Igor XOP Toolkit

Learn More

Igor NIDAQ Tools MX

Learn More