When scientists use a log graph, do they actually convert x/y w/ log?

[hongiddong]

Homework Statement

For example, a linear plot would with x and y values would just plot normally where x and y increases linearly from 0 to infinity.

When we use a log graph, do we just plug in x and y to a graph in which the x and y go from 0, 10, 100, 1000, or do we convert the numbers such as log(x) = new y and then plug it into a graph that goes from 0, 10, 100, 1000?

Homework Equations

The Attempt at a Solution

 

[HallsofIvy]

I'm not sure what your question is. I know it used to be possible, and still may be, to get "semi-log" or "log-log" graph paper. The latter has both axes labeled such that the point labeled "2" is actually at distance "log(2)" from the origin (which is labeled "1"). The "semi-log" paper has one axis like that, the other a regular linear numbering. If you are using such paper, you mark them as they are labeled.

 

[nrqed]

Homework Statement

For example, a linear plot would with x and y values would just plot normally where x and y increases linearly from 0 to infinity.

When we use a log graph, do we just plug in x and y to a graph in which the x and y go from 0, 10, 100, 1000, or do we convert the numbers such as log(x) = new y and then plug it into a graph that goes from 0, 10, 100, 1000?

We do not convert the numbers to log(x) = new value.
We plot directly the initial values x without taking a log. The idea is that the scales on the graph paper are distorted in such a way that the log will be taken into account graphically.


The second option is to use *ordinary* graph paper and plot on it the calculated log values of the x.

 

[hongiddong]

Thank you! I understand it now!

 

[Orodruin]

In the "good ol' times" when you would draw graphs by hand, you would typically have log-log paper with preprinted logarithmic scales. This would make it easy to draw your points in the correct places without having to compute the log. Nowadays most people would simply let a computer draw the graphs for them and it no longer matters much. I have seen examples in scientific papers of people plotting in log scale with logarithmic scales, in which case the values on the scales would not be converted, and of people plotting log10(x) in a linear scale (in which case this would be clearly stated). As long as you are clear about what is being plotted, there should be no possibility for misunderstanding.

 

[Ray Vickson]

I'm not sure what your question is. I know it used to be possible, and still may be, to get "semi-log" or "log-log" graph paper. The latter has both axes labeled such that the point labeled "2" is actually at distance "log(2)" from the origin (which is labeled "1"). The "semi-log" paper has one axis like that, the other a regular linear numbering. If you are using such paper, you mark them as they are labeled.

Probably most office supply stores would not carry log-log or semilog paper anymore, but university stationers might---I haven't checked. If not, there are numerous websites that have such paper downloadable and/or printable.

 

[Chestermiller]

All the graphics packages, like excel and kaleidagraph, allow you to change the scale to logarithmic automatically, so you just plot x vs y, and then, from a drop down menu, switch to logarithmic scales with a click of your mouse.

Chet

 

[NascentOxygen]

Nowadays, there are sites with semilog and log-log paper as a jpeg which you can download and print off. FREE!
You get to choose how many cycles you need.

EDIT. So now I look back and see Ray Vickson already said as much.