CareerCup

This is a huge question and project later on.. Some of the light are here.

1> We need to segregate user based on the geographic location and this will help server to perform analysts properly. (distributed system concepts)
2> In each server, we need to measure each user's activity and update it to the master server in certain scheduling occurrences.
3> Now, we need to use data mining concepts to analyze data.. This again is a master project to provide analytics..

I would rather wish to discuss on this question then type coz it is indeed a massive project

By sending data in real time back to the server . Real time sockets could be used where sockets of both the client and server are always open while the user is on the website .
Getting information like geographical location, latency using ip .
Getting further data using cookies and tallying them every time the user logs on to the website .

This is just what is on the top of my head now .

1) check IP address - geographic info / robots hitting the site / Bad proxy IP hitting the site.
2) check the user agent - to understand the details of what is the user's device info. Are the users using the site on mobile /website, windows or Mac, browser type and version. This tells the user's device info. Also check whether its robots hitting woth malfunctioned useragent or gioglebot/bingbot.
3) looks what pages this IP with this useragent is visiting.
4)

I'm a Java developer and solution is add filter in the project, let that filter capture all the incoming requests

public class LogFilter implements Filter{
public void doFilter(ServletRequest request,
ServletResponse response,
FilterChain chain)
throws java.io.IOException, ServletException {
}
}