数据统计汇总方法（数据统计汇总方法有哪几种）

数据无处不在，没有连接的数据就是一座又一座信息的孤岛，而将这些互联起来，就成了大数据。任何一种有价值的数据都可以作为我们处理的材料，它可以是扁平文件（如WORD文档）、关系数据库文件、也可以是互联网数据库。任何一个数据集经过一些简单的处理和特征提取后，都可以直接进行数据分析和算法的应用，但是首先了解一个数据集是十分重要的，必须首先了解一个数据集的分布状况，属性值的分布状况等等。一般我们采用分布统计汇总方式来做数据状况的描述。这种方法通常从两个角度展开，一种是度量数据的中心趋势，一种是度量数据的离散趋势。

描述数据中心趋势的值有很多，典型的有算数平均值、加权平均值、中位数、众数。而这些数值都各有特征，各有各的特点。比如中位数不受偏大或偏小的数据的影响，因此，中位数适用于对倾斜数据的处理。也有一些我们常用，但是不太在意的数，比如截断均值，其实就是我们常说的去掉最高分和去掉最低分，以此来避免极端数据的影响。

描述数据离中趋势的数值也有很多，比如方差、极差、百分位数，四分位数、和四分位距。这里重点介绍一下四分位数。把数据从小到大排列，并分成四等分，处于三个分割点位值的数就是四分位数。而四分位即第三个四分位数和第一个四分位数的差值。

一组数据集通常可以用5个数来概括数据特征，即中位数、四分位数的第三个（Q3）和第一个（Q1）、最小观测值、最大观测值。我们可以用盒图来对这些数据进行形象的展示。

我们通常用SPSS软件或者MATLAB来绘制盒图。Q1构成盒子的下底边，Q3作为盒子的上底边，盒子里边的红线即为中位数。在盒子外边延伸出来的虚线我们称之为胡须，胡须的两端就是最小观测值和最大观测值，红色的加号表示的数据偏离了数据中心，需要单独处理。

其实，有的数据我们也可以用Excel来进行处理，比如画散点图，我们很快就能够找到数据的离群点。

There are many values that describe the trend of the data center, typically arithmetic average, weighted average, median, and mode. Each of these values has its own characteristics, and each has its own characteristics. For example, the median is not affected by large or small data. Therefore, the median is suitable for processing skewed data. There are also some numbers that we often use but don't care too much about, such as truncating the mean. In fact, we often say that we remove the highest score and the lowest score to avoid the influence of extreme data.

There are also many values describing the trend of data deviation, such as variance, range, percentile, interquartile, and interquartile range. Here is an introduction to the quartiles. Arrange the data from small to large and divide it into quarters. The number at the three division points is the quartile. The quarterback distance is the difference between the third quartile and the first quartile.

A group of data sets can usually use 5 numbers to summarize the data characteristics, that is, the median, the third (Q3) and the first (Q1) of the quartile, the smallest observation value, and the largest observation value. We can use box plots to visualize these data.We usually use SPSS software or MATLAB to draw box plots. Q1 constitutes the bottom and bottom of the box, Q3 is the top and bottom of the box, and the red line inside the box is the median. The dotted line that extends outside the box is called a beard. The two ends of the beard are the minimum observation value and the maximum observation value. The red plus sign indicates that the data deviates from the data center and needs to be processed separately.

In fact, we can also use Excel to process some data, such as drawing a scatter plot, we will soon be able to find outliers in the data.

参考材料： 百度文库 谷歌翻译

本文由learningyard新学苑原创，部分图片来源于网络，如有侵权，请联系删除。