Research On Website Traffic Statistical Analysis Technology

Along with the rapid development of the Internet technology, both of the Internet usernumber and the websites number increase sharply, at the same time, a lot of website trafficstatistical analysis system tools emerged in the Internet field. Through the comparison of thedomestic mainstream website traffic statistical analysis system tools, the author found that allof them could not give out accurate statistics report about the unique visitors within a periodof days. This paper puts forward a quick calculation method of calculating the number ofunique visitors. This paper is focus to realize the website traffic statistical system, especiallycalculate out how many unique visitors visit the website in some days based on the researchof three kinds of website traffic analysis technology and the research of distribution of thewebsites visitor numbers. Finally, it is given out the algorithm implementation code processand test results. The author’s main task is to do preliminary research of distribution of thewebsites visitor numbers and finish website traffic statistical system which include datacollection module, data analysis module, data storage module, data extract module and so on.In order to count out the unique visitors number，the website analysis system use themethod of combing Log File analysis with Page Tagging which can accurately identify thevisitors by Cookies. It also builds a cache field and using NoSql database for a rapid inquiresto visitors to speed up the system speed in the data analysis module. Finally, website trafficstatistical system is completed. Users can operate the website analysis system through specificprotocol.The realization of the statistical algorithms make website traffic statistical analysissystem can quickly calculate out the number of unique visitors who visit the website in somedays, which function most of web analysis tools don’t have. At the same time, it makes thewebsite traffic statistical analysis system product to be more competitive after it waspublished in the web market. On the other hand, the proposed algorithm in this paper alsoapplicable to the statistics about the number of login users or the number of IP that visit thewebsite in some day.