A5 (www.admin5.com) stationmaster net daily recommended: I believe that many of the webmaster friends may be in the analysis of website operation and site data problems and distress for web site operators, data analysis has a certain important position, how to carry on the data analysis, such as He Dingyi, the loss of users how to do data analysis and so on a series of the problem. Below sorted out some of the site operations and data analysis of the actual combat experience, I hope to help some webmasters solve problems.
1 how to reasonably define user churn
did not update the blog for a long time, this article to write some about the user loss content. Active users before the release of the site and the loss of the user of this article on the website of the active users, the loss of loss of customers and new customers are defined here, corrected for the loss of the user name English, generally for the loss of the user commonly used English as "churn user", before the use of wastage, away and lost are not too specification. Later a friend asked to do correlation analysis of loss of the length of time the loss of the user’s choice in the end how much is reasonable, especially after the "real" site analysis of this book, I have mentioned the length of time how to more accurately define the loss inside, may explain the relatively simple, there are still problems in this area friend message feedback, so here with an article to explain.
article read: site data analysis: how to reasonably define the user loss
2 data from users to talk about data analysis this thing
regardless of doing statistics or look at other people’s data, the first step is always the reliability of data acquisition. If the sampling data, we must look at the sampling method to see what kind of error may exist. If it is their own data, but also to see whether the data acquisition itself is scientific, such as statistical user behavior is generally used JS callback, if you do not use Apache log statistics, the results would not fly.
: reading data from the user data analysis on this matter
3 how to conduct data analysis
webmaster circles spread the word: "normal content, 2B webmaster webmaster do outside the chain, Niubi webmaster do data", if you are from the 2B off into a niubable hope that this article can wish you a helping hand, of course, in addition to your own effort and learning.
article read: hand taught you how to conduct data analysis
4 Web site data analysis data warehouse related issues
read Inmon’s "build data warehouse" and "DW 2" before, and another data warehouse master Kimball "data warehouse"