Hi Mehwish,

It is difficult tto add something to the many interesting suggestions already given here. I will give it a try anyway ! On the R-project site you have a searchable listing of books related to R that are useful to the R user community:

http://www.r-project.org/doc/bib/R-jabref.html

Another nice piece of work is the R gallery:

http://gallery.r-enthusiasts.com/

You will find lots of graphs (with the code needed to generate them) that you may wish to use and modify for your practice. I did this when I started with R and found it useful.

I also used to consult this R-tips page that gives very good suggestions to some practical issues and questions:

http://pj.freefaculty.org/R/Rtips.html

Hope this helps,

Tito

I am afraid too many opinions might confuse u.

I am not sure whether your question is about statistical package R or correlations (r). As regards R for windows or Mac Francois answer above covers it all. However if you are asking about correlations..here I go...

I am a PBL teacher and I start with problem.

I suggest following steps:

1) Open SPSS

2) Go to variable view and make a few (say 4 Numerical) and 4 string variables

3) Now enter data intelligently. e.g if your Numerical variables are serum insulin, random BG, fasting blood glucose, and HBA1c. enter them in a way that increase insulin decreases glucose and decreased insulin increases it.

4) Like wise e.g string data can be 'number of children, values 1,2,3,4,5 or more, psychiatric problems in family, IQ of children...and so on, enter the values properly.

5) Now go to analyze in the menubar, drop down and click on 'correlate' and then 'bivariate' in the window that opens. Shift the numerical variables that u want to correlate to the right side coloumn, select pearson from below and OK...take a sip of tea...in 1 minute (it is a bit slow) u will have 'r' values as well as significance of correlation. if there is a minus sign with 'r' value this means correlation is negative, i.e e.g if insulin increases glucose will decrease. however there will be no sign if correlation is positive e.g if distance covered by train increases, fare will also go up. If the 'r' values are more near to 1.0 the correlation is more linear.

6) For categorical (string) data select Spearmann's and the rest of the story is same..

If you want to read more, do not go here and there, just go into tutorial of SPPS or statistcial coach.both will be found in Help of SPSS.

Bless u

Initially I invested in Crawley's "R Book" -- which I still like enough to order the 2nd edition. You may also want to look at www.jstatsoft.org, the free online Journal of Statistical Software, which has a lot of R coverage and tutorials in each article describing a new package.

Thank you Francois: I found the article interested, though it covered data management only. The R manual is also helpful but I get lost when go for advanced analyses.

Thank you Wangan! the statstudio.net is interesting. I bookmarked this website for Quick R help. Though, the others require bit time to understand.

Badar: Correlation and Regression analysis is my area of expertise. I am a statistician basically. I wrote, I frequently use all other statistical software except R and wanted to get expertise on the same software.

Mitchel: Is Crawley's R book available online?

Khurram: Yes, I found statmethods.net interested and bookmarked it.

Well, which option do you use in Rstudio to proceed? I found three and found one relevant: :"Open R Project", but still confused.

Good, interested. Thank you both wangen and khurram! I will really try this.

Mehwish: the Crawley book is only Hardcover. I also liked Brian Everett's Handbook of Statistical Analyses Using R, which is 2nd edition and either paperback or Kindle.

thanks, I have downloaded it. Well, it has too weighty text. Let me experience it too.

Thank you Jose.

I will suggest to start playing around with basic functions in R (probably you have done that enough by now).

The next step is to learn writing your own functions. For example, write a function that compute correlations. Then compare the results with those from cor() function in R. Find mistakes, fix them until u get it right. Experience through mistakes is the key to learn R. I swear, this method will amplify your confidence (or rather addict?).

Eventually, try to include arguments, loops etc.

Interesting Suggestion! I will do this during weekends.

I just started using R and I'm really loving it. I also am using R Studio and find that interface a great one to use. One thing I did find was installing packages from R studio did not work, I needed to install via the basic R package. (Do others find this as well? I saw there was an issue with this.)

That said, the way I found confidence is to just decide to analyze the next dataset in R. I also realized that saving my scripts was an important step, then in R studio highlighting them and running them, but perhaps with your previous experience you might have already known that.

I have also found youtube videos that demonstrate specific techniques for various analyses I want to run in R helpful, like prcomp() for principal component analysis using SVD, etc. I find that the screencasts of people working in R not only show how to do analyze data in R, but also as a byproduct, you also get to see their workflow in R and how they organize their scripts, windows, etc. for efficiency.

I'm looking forward to communicating with this R community here at researchgate as well. :)

Hi Mehwish,

It is difficult tto add something to the many interesting suggestions already given here. I will give it a try anyway ! On the R-project site you have a searchable listing of books related to R that are useful to the R user community:

http://www.r-project.org/doc/bib/R-jabref.html

Another nice piece of work is the R gallery:

http://gallery.r-enthusiasts.com/

You will find lots of graphs (with the code needed to generate them) that you may wish to use and modify for your practice. I did this when I started with R and found it useful.

I also used to consult this R-tips page that gives very good suggestions to some practical issues and questions:

http://pj.freefaculty.org/R/Rtips.html

Hope this helps,

Tito

Thank you both of you!

I don't think I'm going to use R to manage large data sets anytime soon. R is nice, though, and I call R within SAS- See, for example: http://support.sas.com/rnd/app/studio/statr.pdf

R is far from being user friendly...

how to get experience? for any beginer I would advise Rcomander

install.packages('Rcmdr')

library(Rcmdr)

Then try to do some analysis, graph... use some plug-in, it helps for beginers bc it make by itself the comand line to used in an user-friendly way.....

When you will realize it isn't good comands and that you should better use any other way... then you will be an experienced user :p

Really, interesting.

Thank you!

I think R is very useful and interesting for statistics practitioners. I am sure that Google will help you better.

Hello Mehwish,

The following link will be very useful for you to embark in R:

http://www.r-tutor.com/

There you have everything you required plus give you examples as well.

Kind regards.

Kumar

Kibria! I tried the google earlier and browsed it much. I got many complex documents, therefore, I put question here. Well, Kumar, thats another good site. Thank you!

Also, I've used Quick-R quite a bit as well: http://www.statmethods.net/

Great Wicher! really valuable. Thanks

Hi Mehwish, there is a site cousera.org...it holds courses and R course is going to start on 28th jan..you can register yourself and get free video lectures and assignments and also a certificate in some cases..have a look..really helpful site..

Thanks Sonal for informing me this. I visited coursera earlier for searching other coursework. Well, this suggestion is also pretty well that one can remain in bound practicing the software consistently.

The samples that come with R are very illustrative