CODEWITHSUNDEEP
rdata-visualizationtrend
LoomyBear
LoomyBear

Reputation: 449

How to build a trendline on a graph in R

I've checked everywhere, and people refer to examples that I can't understand (yes I'm kinda slow). Could anyone please explain me how to build a logarithmic trendline in R?

Here's the working example:

myds <- c(23.0415,13.1965,10.4110,12.2560,9.5910,10.7160,9.9665,8.5845,8.9855,8.8920,10.3425,9.3820,9.0860,9.6870,8.5635,9.0755,8.5960,7.9485,8.3235,8.1910)
plot(myds)

I can't find a simple way to apply regression trendlines. I'm interested in particular in the logarithmic and the linear trendlines. Is it possible to do without connecting any new packages?

Good sirs, please be kind to clarify!

Upvotes: 6

Views: 72959

Answers (2)

Daniel
Daniel

Reputation: 7832

Since you have not provided reproducible examples, I'll post some links, which I think might help you:

An example for a simple, linear trend line is here: http://www.r-tutor.com/elementary-statistics/quantitative-data/scatter-plot

Furthermore, there has been a thread on this over at SO: How do I add different trend lines in R?

Using ggplot would make it a bit easier, as you can use the smooth functions.

Upvotes: 2

coulminer
coulminer

Reputation: 368

since you had some data points missing, I took what you had provided: the six points.

edit - now the full example was available

  1. A trendline is just a regression, and regressions are run most simple way like this: a<-lm(outcome~predictor) -- in this example the object a will hold your regression parameters. To get the values of your new trendline model, just use predict(model_name), or in your case predict(a)

  2. Adding line to a plot is dead simple. Just say lines(b), where b specifies the line you want to plot after you have used the plot() function.

To wrap it up:

[![myds <- c(23.0415,13.1965,10.4110,12.2560,9.5910,10.7160,9.9665,8.5845,8.9855,8.8920,10.3425,9.3820,9.0860,9.6870,8.5635,9.0755,8.5960,7.9485,8.3235,8.1910)
x <- (1:length(myds))
plot(myds)

#make the main plot
plot(x,myds,ylim=c(5,30),xlim=c(0,20))

#add linear trend
lines(predict(lm(myds~x)),col='green')

#one more trend
lines(predict(lm(myds~log(x))),col='red')][1]][1]

resulting plot image

Upvotes: 7

Related Questions