The Student Room Group

What is the gradient of a logarithmic scaled graph???

Help!

How do find the gradient from a graph that has logarithmic scales?

y = mx + c obviously does not work or give the correct trend, so is there a general formula for this?

I need to find the gradient from some graphs (on paper) and then somehow transfer them onto Excel to create as similar a graph as possible - the lines are straight on the paper-form graphs (with the log scales on). But when I read points from the graph on paper and plotted them in Excel, the line was not a directly proportional line and not entirely straight, as the graph on paper is.

Help help help

Katy katy katy :smile:
Reply 1
linka1
Help!

How do find the gradient from a graph that has logarithmic scales?

y = mx + c obviously does not work or give the correct trend, so is there a general formula for this?

I need to find the gradient from some graphs (on paper) and then somehow transfer them onto Excel to create as similar a graph as possible - the lines are straight on the paper-form graphs (with the log scales on). But when I read points from the graph on paper and plotted them in Excel, the line was not a directly proportional line and not entirely straight, as the graph on paper is.

Help help help

Katy katy katy :smile:


Can you upload your excel document as an attachment...then perhaps I can help.
Reply 2
If you want to create a graph with a logarithmic line of best fit on Excel, here's how you do it.

Produce the graph of your values as a scatter graph, not a line graph, with no line of best fit. Single-click on a point on the graph and it should highlight all the points on the graph. Then right-click, keeping your cursor above the point on the graph you clicked on originally, and select 'Add Trendline...'. Once you do this, a window called 'Add Trendline' will appear. Under the 'Type' tab, select 'Logarithmic' and click 'OK', and Excel will automatically produce a logarithmic line of best fit for you for your graphs.

Hope that helps. :smile:
Reply 3
Thank you!

I gave up in the end and read the data from the graphs instead. Took longer but at least I wasn't wasting time trying something that might not work.

Katy :biggrin:
Reply 4
lny = mlnx + c
y = e^(mlnx + c) = (e^c)(x^m)

(1/y)dy/dx = m/x
dy/dx = ym/x = (m/x)(e^c)(x^m)