Difference in Line Graph and Scatter Plot

v9991

Trusted Information Resource
i tried to fit a data in both line graph and scatter plot.
then generated a trend line., but both of them are giving different values for slope, r-sq, etc.,
why do i see such difference?
 

Attachments

  • book01.docx
    13 KB · Views: 111
T

t.PoN

the line graph is usually used for non-numeric x-values.
the scatter graph is used for numeric x values (it uses linear regression). the difference is in calculating the error.

To explain it:
the line graph will try to fit each point on the line.
but the scatter will try to find a correlation between the whole set of data (you may find a point outside the line but the correlation will fit better for the whole set of points.

i suggest you stick with scatter formula
 
Last edited by a moderator:

Bev D

Heretical Statistician
Leader
Super Moderator
May I ask what you are trying to understand with the graph?
 

v9991

Trusted Information Resource
sorry, I missed to mention that, we were trying to assess the linearity. and assess the R-sq. as an acceptance criteria.
 

Bev D

Heretical Statistician
Leader
Super Moderator
Linearity of what? What are you measuring? what is th input (x axis value) and the output (Y axis value)
also you have very few data points so the R squared value may be artificially way to high...

I assume the line chart uses the 'trend line' option? in any case I would never that use option for anything...

I woule go with the scatter plot / standard linear regression approach at least to start. but you really need more data
 

J0anne

Joanne
You need to choose the correct graph

If you want to show data trends, with only one set of numeric data, then choose the line graph. The line graph uses values on the vertical axis against categories on the horizontal


If you want to show a comparison of values, then choose the scatter graph. This graph shows values on both axes
 

v9991

Trusted Information Resource
Linearity of what? What are you measuring? what is th input (x axis value) and the output (Y axis value)
also you have very few data points so the R squared value may be artificially way to high...

I assume the line chart uses the 'trend line' option? in any case I would never that use option for anything...

I woule go with the scatter plot / standard linear regression approach at least to start. but you really need more data
i understand/understood the application for specific purposes, BUT, i had a academic/curious question, that why is there such a difference, for same set of data. what/how does excel give two different formula?

is it that, excel treats x-axis of line graph as some coded values 1-2-3-4-etc.,
and in scatter plot, it actually accounts for exact values of the graph variables.
 

Miner

Forum Moderator
Leader
Admin
is it that, excel treats x-axis of line graph as some coded values 1-2-3-4-etc.,
and in scatter plot, it actually accounts for exact values of the graph variables.
You hit upon the correct answer. Substitute the integers 1 through 6 into the formula for the line graph, and the equation provides the corresponding y-values on the line graph.
 
Top Bottom