Self-Study Plan for Becoming Statistical Analyst? Can any one please suggest me a self-study plan for becoming a Statistical Analyst? I need suggestion something similar to Self Study Plan for Quantitative Analyst
 A: I am currently in the process of doing this (sort of), and think I have had moderate to great success. I have relied solely on primary literature and experience without any textbooks. Here is what I found most helpful:
1) Learn R. I don't see how anyone can get an intuitive understanding without running many, many simulations. Perhaps with a strong math background this is not as necessary.
2) Learn the history of statistics and the source of various controversies, for example bayesian vs frequentist (Neyman Pearson, Error Statistics) vs fisher (inductivist) vs likelihoodist vs the NHST (Null Hypothesis Significance Testing) hybrid. There may even be more! Especially learn about the history of the NHST hybrid controversy, I had no idea how bad it was and was taught that as "Statistics". 
3) Get data of the type you think you will analyze (timeseries, comparing groups, etc) and analyze it in many different ways to learn the strength and weaknesses of different approaches.
A: I'll answer from the perspective of an economist who did lots of econometrics.
If you haven't done any R/Python/Matlab etc. and you want to be able to use statistical techniques, understand when to use which, how to apply them I'd suggest you start with Chris Brooks' Introductory Econometrics for Finance. The book uses EViews for all examples, and it's great for a beginner as it explains both the theory, gives an example, and then shows you how to do it step by step. I loved it!
If you want something more interactive, and with programming head over to coursera.org and check out their Stats 101. I think it uses R, and will have plenty of reference material. 
A: I am a college dropout and have done this successfully. So follow the steps i enlist below :


*

*Learn R or Python.

*Try to understand what a probability distribution is.

*Learn about the Normal, the binomial and the uniform distribution. Consider only a single variable.

*Find out as many situation where the above three distributions have been applied and find out as many properties about these as you can.

*Try and visualize two variables, both normally distributed. in one case they are positively correlated and the other case they are negatively correlated.

*Try to plot different variables and try to find differences between them using just what you see in the plots.

*Once you have done this, take up any dataset(s) and start finding things about it using the distributions and the plotting.

*generalize your knowledge to other procedures like regression.

*generalize your knowledge to other distributions.


Some books to help you out :


*

*http://www.amazon.com/Statistics-An-Introduction-using-R/dp/0470022981

*http://www.amazon.com/Using-Introductory-Statistics-Chapman-Series/dp/1584884509/ref=pd_sim_b_5

*http://health.adelaide.edu.au/psychology/ccs/teaching/lsr/
Finally start analyzing something that you like using the techniques that you have learned. Now you are ready to go to the advanced level.
A: Learn R. Absolutely, definitely do so. R is open source and I think will soon become the de facto standard for the statistics & machine learning. You also need to learn a scripting language and I would strongly recommend Perl. It doesn't have the fanciness of other languages but it is super powerful and guaranteed to be a solid return on invested time, not to mention the vast swath of packages that are already created for you on CPAN. Finally, you also need to get your hands on some books to learn statistics. Some of the best books in statistics, can be found here. Linear algebra is also a good to know but not necessary. If interested here is a list of books for learning linear algebra. HTH
