Doing statistical analysis and charting with large data? I have 4301 lines of data from my science project.  I have to do unorthodox statistical analysis for my results because they are 4-dimensional; I have three independent variables.
With Excel I can do a lot of analysis, even multivariable linear regression, but what I need is a step up:  Multivariable nonlinear regression.
I have R.  Is there a way I can do it with R?
Also, is there a way to generate multiple 3-dimensional graphs with slices of the data automatically?
EDIT:  I've searched with no luck.  Is there any way of doing multivariable nonlinear regression automatically?
 A: It's not real clear what you're asking.  Is the following useful?  I knocked it down from 4301 to 430 lines in case your machine is slow.  You can move the little green box around to see how those points fit in all of the other graphs.
library(TeachingDemos)

#Build 3 independent variables (for speed reasons, I knocked it down to 430)
x <- rnorm(430)
y <- runif(430)
z <- rnorm(430)

#Build the dependent variable
dep <- 2 + (0.5 * x) + (0.9 * y) + (-0.8 * z)

#Put everything in a data frame
df <- data.frame(x=x, y=y, z=z, dep=dep)

#Plot it and look at chunks of the data
dra <- tkBrush(df)


A: What sort of nonlinear model do you want?
4300 lines of data is not very large in today's world.  Data sets containing millions of records exist.  Certainly 4300 will be no problem
This link to a work by John Fox may be helpful
http://cran.r-project.org/doc/contrib/Fox-Companion/appendix-nonlinear-regression.pdf
it highlights use of the nls function in R, which is part of the nls library
Hope this helps
A: Save your table as a csv file, with one header row that contains the names of your 4 variables.  Save the following code as an R script file in the same directory as your data, and then run it.  Lets say your variables are called X1, X2, X3, and Y, where X1, X2, X3 are your independent variables and Y is your dependent variable.
MyData <- read.csv('path/to/MyData.csv')
model <- lm(Y~X1+X2+X3,data=MyData)
predictions <- predict(model)

model
plot(model)

Building a linear regression model and diagnosing it is relatively simple in R, even with 4000+ lines of data.  What exactly are you trying to do with this dataset?
You can build a non-linear model using the "loess" or "glm" commands, but specifying and diagnosing such models is more difficult than simple linear models.  Again, if you can be a bit clearer about what exactly you need to do, we can help you more.
Fitting a linear model with 3 dependent variables and 4000 observations isn't exactly unorthodox, and you can do it quickly and easily in R.
A: From your comment to my previous answer "....I also want to automatically generate 3-dimensional graphs of each slice of the three IVs (as in leaving one IV constant while using the other two as x and y, and the DV as z)....."  Try the following (using the same data as above):
library(lattice)

#Build 3 independent variables
x <- rnorm(4300)
y <- runif(4300)
z <- rnorm(4300)

#Build the dependent variable
dep <- 2 + (0.5 * x) + (0.9 * y) + (-0.8 * z)

#Put everything in a data frame
df <- data.frame(x=x, y=y, z=z, dep=dep)

#Break up the x data into 4 chunks of equal counts
df$x4 <- equal.count(df$x, 4)
cloud(dep ~ y + z | x4, data = df,
      zlim = rev(range(df$dep)),
      screen = list(z = 130, x = -65), panel.aspect = 0.75,
      xlab = "y", ylab = "z", zlab = "dep")

#Break up the z data into 2 chunks of equal counts
df$z2 <- equal.count(df$z, 2)
cloud(dep ~ x + y | z2, data = df,
      zlim = rev(range(df$dep)),
      screen = list(z = 160, x = -80), panel.aspect = 0.75,
      xlab = "x", ylab = "y", zlab = "dep")

The above generates two plots.   Below is the first plot.

