Blue skies and statistics in Beijing

Intermediary transaction http://www.aliyun.com/zixun/aggregation/6858.html ">seo diagnose Taobao guest cloud host technology Hall

I am a Beijing person, from small to large has not left Beijing, Beijing, kindergarten, primary school, middle School, University, Beijing work, Beijing marriage. If there is no feelings for Beijing that is impossible, do not like this piece of land I can still like where? But this year Beijing's spring is really not very much, the sky is always gray, no translucent. There has been no precipitation since winter, and nothing has been done to get angry. The temperature rises very quickly, but does not have the bright spring, like is to be covered up steamed. When the two sessions, the national attention, Beijing citizens should conform to the good situation of joy is palpable. After a few months, the greatest Olympic Games in history are about to begin, should Beijing's citizens be excited, yearning and proud? But it seems not so, at least I do not have such a heartfelt joy.

Well, the first paragraph is over, the mood of the people of Beijing in the spring of 2008 is not the point I want to say, I want to say something about statistics. The last time I read the Zac written statistics of misreading and misleading, deep thought. At the same time think of the recent "Phoenix Weekly" on the Beijing Blue Sky Data article, from Beijing after the successful Olympic Games, the state media has been repeatedly fooled Beijing's blue sky has increased how many days, and how many days in advance to reach the beginning of the plan ... And actually? Phoenix Weekly explained that, in addition to the Government's great efforts in environmental protection, blue sky data on the statistics made two points adjustment: the first adjustment of the sampling point location, reduce the urban collection point, increased the suburban collection point; the second adjusted the statistical index, removed the not easy to meet the sulfur dioxide replaced by the easier dioxide. This is enough to show how data statistics can be manipulated.

The above text is written yesterday, Beijing has finally seen the sun and Blue sky, Zac also wrote a "statistics in the Limits of SEO", then I also add a case to say the problem of statistics. Some days ago the company's internal FTP server stability frequently complained about, so I have some statistical analysis of server logs.

First, the monthly statistics on the upload success rate and upload speed, found that month, February, March data no difference, so the heart is very angrily, thinking that the data will not lie, clearly the quality of service no change, why also complain!

is looking for everyone theory, suddenly more a heart son, and daily statistics upload speed and success rate, found the problem, although the recent total success rate has not changed, but January most of the day success rate is very high, only a few days of very low data, pull down the success rate. and February or March is the average daily success rate is low, no wonder we have to complain ...

So how do you find the cause of the problem? We didn't do anything. How did the quality of the FTP drop? Another statistic was made: statistics of transmission speed and success rate by users. The results came out and found: The original is that a few people complained about the total error, accounted for the total number of errors more than 80%! And most of the silence is actually ftp very good. After looking for these people specifically understand what the client is, how to set up, home network is what ... found that one of the most failed colleagues was using power broadband ... That's no wonder.

What do you see in the twists process? It's easy to talk with data.

From: Li Yun-yang (Reggie Lee) blog, the Blue sky and statistical data in Beijing