Some of my knowledge of website statistics

Statistics

I have some knowledge of the website statistics, hope to be helpful to everybody. Everyone take a good look at Ah, and then add.

Web counter concept, as the name suggests, it is used to record the number of visitors to the Web page.
Generally speaking, counter has three kinds of forms:
One is by the webmaster through some JS code, you can directly display the number of visitors on the page; This method is more old-fashioned, in addition to the simple display of data, no other role.
Alternatively, the log analysis software installed on the Web server is used to implement statistics by analyzing the server's logging directly. The biggest advantage of this analysis is generally very accurate, but the weakness is also obvious, first of all, must make their own server to install, loading the speed of the server does not say, once someone else this software left a backdoor, your server will die. The function of this statistic system is more abundant.
The third method by embedding the third party counting web site of a section of JS code as an HTML logo, by the statistical site to collect the visitors information statistics analysis and calculation. Most of the current sites have taken such a counting method, where we only analyze such counters. This statistic is more powerful, especially if it provides a lot of useful information.

  Generally speaking, the code placed in the Web page will contain the server address of the statistics server and the user name of the person being counted, each one who visits this webpage, will read this JS code, and then establish the link with the server of the statistic website, and carry on the unified analysis statistic to this information.
based on such a mechanism, it is decided that such a statistical form would have the following characteristics:
     faster: equivalent to adding a link to the statistics server in the Web page, the speed of statistical server is generally very fast, As long as the Web site of the server and statistics servers are in the telecommunications or Netcom network, visitors basically do not feel the load time. But if the telecommunications and netcom across the network, or a slight delay.
     statistics are more accurate: as a result of a unified analysis of the data collected on a third-party server, there are certainly no two other methods for obtaining accurate data. But if the statistics service configuration is relatively high, the network bandwidth is relatively large, and can guarantee the visitor statistic code loading integrity, basically the error is very small.
     has a lot of statistical functions: the browser in the link with the statistical site, according to the current Web site Interconnection protocol mechanism, the statistics site's server can get a lot of information about browser settings and access to records. This information generally includes: the IP address of the visitor, the browser type, version, installed plug-in information used by the viewer, the screen resolution used by the viewer, the color depth used, the system default language set, the time zone in which the viewer is located, and the last access address of the person. Through the collection of this information, and a certain number of algorithms for analysis and summary, statistical sites can get a lot of information about the target, they will be collated after feedback to the webmaster, it formed our daily see the site statistics system.
Generally, such a Web site system is written in a very popular ASP or PHP language, and is configured based on a SQL or ms-access database.

Has the following basic functions:
At the very least, you can count the number of visitors ' Independent website address (IP) and the amount of refresh (PV) of the Web page, and can observe the results according to the hours, days and months. Some of the more advanced statistical systems can also be able to identify independent computers, more accurate than IP.
You can record the source of the visitor, which is the last site visited by the visitor's browser. and analysis of these sources.
If the source is from a search engine, you can identify the type of search engine and use the reverse coding technique to learn the keywords used by the caller when searching.
Other information about the visitor, including the location of the visitor by analyzing the visitor's IP address, the browser settings of the visitor whose system information was obtained, and so on.

Hardware and software features:
Generally speaking, do an online statistical system, need to invest a lot of money and energy, statistical site different other sites, than the need to have a lot of computing power, large memory, storage space reorganization of a very strong server can be achieved. According to my observation, the situation of statistics server is divided into independent server or server group Two kinds of forms, the former situation for a single server where to accept the user registration, statistical calculation, each can withstand a certain amount of traffic impact, one to reach the limit, to change another; the latter connects all servers into one group, sharing bandwidth , in parallel together to accept registration, statistical calculation. Obviously, the latter of the parallel server group structure than the former independent server more advanced, everyone in the registration counter should choose such a statistical system.
Domestic counter statistical software is generally purchased from abroad, by the statistical site after the installation of the Chinese, there are also some of the shared ASP statistical procedures to provide statistical system webmaster download and install the use. So we often see the same or almost identical statistical system on the Internet. Such a statistical system because the core software is not self-developed, the statistical site has almost no independent provincial, expansion of the capacity. Register counters, should pay attention to the choice of statistics website has the independent copyright of counters, and constantly introduce new features of the statistical system. Their technical strength can be trusted, and will not simply rely on the * forced pop-up plug-ins in the form of profit.

Profit Model:
Are these statistical services provided free of charge? We can see that there is no mention of charges in the registration notes for most statistical sites. But there is no free lunch, and the statistics sites need a server to run, and they need a lot of spending to keep them going. How do they make a profit? Basically, the statistical site is through the implementation of the statistics in the JS code embedded pop-up ads or plug-ins to achieve profitability, such as Baidu and 3721 of Plug-ins, so that is divided. But to accept the statistics code webmaster, this practice is very cruel, every forced pop-up plug-ins, netizens on the site's good impression will lose a little bit. This pop-up plugin is embedded in the download to the local statistics code, because this code can be changed at any time, almost impossible to be blocked. Statistics site can be selected for any franchisee, arbitrary settings pop-up content and ratio. Similarly, in addition to Plug-ins, they can also set pop-up ads windows, Web pages, file downloads, or even direct jump to other places or even more malicious code and even viruses. Just imagine if a virus or illegal content pops up, and your site is shut down basically. When registering counters, we must try our best to select the statistical system without plug-ins.

To sum up, the online statistics system to achieve, the key lies in the statistical provider server quality and statistical code efficiency. As long as the two points are done, we can provide fast, stable, functional and comprehensive statistical system. In addition, whether to include pop-up plug-ins, but also to assess the quality of a statistical system is a decisive parameter.