You'd need to explain your model issue first for there are different threads for a proper answer at least. IMHO, initially, and most important, there is the system modeling aspect. It will drive/define the better software that should be used. Then you'll have the data collection process: what data, its meaning (influences modeling), sample size, etc. With the collected data sample, one must perform statistical analysis methods to know its characteristics and learn about the general statistical behavior of the system that shall be simulated. Hope this helps.