Now, data analysis is not only a basis for decision making executives and middle managers need, but also become one of the essential ability of ordinary business people, such as HR field, "with the data speak" the same rules in the area of human resources Be applicable.
In recent years, major companies popular job: HR Analytics , is a good proof, many companies paying jobs, human resources jobs data analysis, but it is still a lantern could not find the right people.
The reason is simple, HR data analysis, HR is one of the most scarce capacity, the number of HR data analysis, and finally turned into a listed number. Template most common so-called human resources data analysis, might be this:
Staff gender, work experience, geographical distribution, education level, etc., the list is good data, coupled with a simple chart.
However, such data, master the basic ability of people to use Excel can easily figured out, why should we hire well-paid HR data analyst to do? Such numbers are tallied, and in the end what it represents? Good or not? The meaning behind the sex ratio, the boss is to guess it?
Only numbers listed, there is no analysis of such data not worth mentioning. And all of the data analysis, there should be value-driven business.
HR data analysis, data should be obtained by relevant organizations and personnel, to business problems in the organization and possible future talent and are forecasting, early warning, anticipation, and put forward the views of decision-making to the business.
Here we are in a real case to explain valuable business analytics How to Make.
First, find and analyze business pain points
Wang is the commissioner of human resources of a company headquarters, every month, every quarter, every year, Wang should be responsible for collating new employees during this time period, turnover, mobilization, and staff for the constituent features, KPI performance , payroll and other management.
However, the company's employee information stored in the OA system, and salary information and stored in another set of cloud ERP systems, some branches have their own business systems. Not get through between these data, like one island, which together requires a lot of manpower, and even took the boss Jin collated data, which also lag, accuracy and other issues.
such as:
1、小王每个月月中就要发布通知，让各分公司将本月的人力资源信息开始汇总，之后进行上报。从发布通知到上报结束，往往需要经历2周甚至更多的时间， 白白耗费了许多人力不说，还存在严重的滞后性。
2、人力资源数据采用Excel进行上报，绩效、薪资的计算也是人工使用Excel进行核算，篡改、纰漏等数据准确性问题无法把关，全靠业务人员的责任心和专业性。
3、数据以月份为单位进行存放，当需要汇总观察趋势时候，就需要将N张Excel进行整合，费时费力，Excel的性能也堪忧。
因为以上的种种数据问题，小王每个月的大部分工作时间都在整理、核算数据，更不要提什么数据分析了。
二、选择分析工具，进行数据处理
为了解决上述问题，小王需要的是一套能够实现多数据源整合+数据处理+数据分析的软件。小王的这一需求，是许多企业推进数字化进程中的共性需求，估计看到这里很多人就想到了商业智能（BI）分析工具。
没错，BI工具能够满足从数据源到业务人员数据分析的完整流程，下面我们就用国内BI市场占有率第一的数据分析工具——FineBI，来看一看如何利用BI工具解决小王遇到的数据难题。
1、数据连接
如何将不同来源的数据基于业务分析需求进行分类管理呢？
以FineBI为例，通过多源的数据连接就可以将数据进行整合，比如传统关系型数据库MySql、Oracle和Kylin，Hive，Spark等大数据平台，从而打破业务系统中的数据孤岛。
如果选择直连模式，则可以实现实时数据展示，上一秒发生的人员变动信息，下一秒就可以在FineBI中展现出来，消灭数据迟滞的同时也降低了人力浪费，大大提升了决策效率。
2、数据关联
在数据连接的基础上，将数据表添加至FineBI的业务包进行管理后，管理员可以给这些数据表添加对应的表间关系。
- 对于IT管理员而言：只需要配置基础的数据关联和权限，分析用户无论如何进行数据处理，都一定是在其权限范围内操作，而且自助数据集的关联也可以自动继承，不需要管理员再进行配置
- 而对于业务分析人员来说：分析用户可以拿到自己权限范围内所需要的数据，进行无限次处理和分析
图为本人力资源案例的表间关联。
3、数据加工
在设置好数据连接，添加好需要的数据表并设置了表间关联后，我们可能还需要对数据进行一些处理，比如过滤掉一些不需要的数据，或者进行一些计算等等。
这时候就可以使用FineBI的自助数据集进行数据的二次加工，进行比如过滤、分组汇总、新增列、排序、合并、挖掘、R语言分析等数据处理功能，可以让不懂代码的业务人员也可以快速上手，轻松搭建各类分析模型。
4、自助分析
数据治理这些后台的底层工作完成后，业务人员就可以针对数据进行探索式分析了。直观的拖放式页面，降低业务人员学习的成本。而联动、钻取等OLAP分析功能，更帮助业务人员从多个角度探索，深入挖掘数据背后的价值。
图为FineBI的数据分析组件页面，只需要鼠标点点，便可完成复杂的可视化分析：
三、输出可视化结果，比对分析
1、新入职员工分析
我们可以通过探索分析发现以下特点：
- 今年每个月雇佣的员工数都比去年多， 有几个月的雇佣员工数明显很多；
- 从按地区和种族划分的新员工计数和在职员工计数组合图可以看出，我们在东部区域招聘的员工较之更少；
- 按年龄组划分的新员工增长率变化瀑布图显示，我们主要招聘的是年轻员工。 这种趋势可能是因大部分工作都是兼职性质所致；
- 新员工数(按性别) 饼图显示，新员工数按性别大致均分；
2、离职员工分析
我们可以通过探索分析发现以下特点：
- 左边的两个组合图显示了，与去年同期相比在职员工数和离职员工数的变化情况。 今年的在职员工数比去年多，这是由于快速雇佣所致，同时离职员工数也比去年多。
- 8 月的离职员工数比其他几个月都多。通过点击不同的区域、性别进行探索分析，发现其中东部地区的离职人数明显更多。
We note that in a pie chart, the number of serving employees by sex and age group sharing. Then you can select different age groups of all ages to know whether gender-sharing.
3, poor staff analysis
The last part is to explore the difference between employees. It is defined as the difference between the employee not to work more than 60 days would leave employees. We are quick to hire, but if we hire outstanding candidates for employees?
Select the left side of the "area" filter assembly "Northwest" and click the "difference in the number of employees (by gender)" donut diagram "male." Research "difference in the number of employees" Other chart on page. We note that the difference between the number of employees to be more male than female, and group A, there are many poor people.
4, poor northwest region Number of employees
If the study "difference in the number of employees (by gender)" circular chart, and select a different area of the filter assembly "area", it will be noted that the eastern region is the only difference between the number of women employees than men and more areas.
Finally, aside tutorials, real time
The above content to show you a simple analysis of a case of human resources, including basic data analysis process, but want to really grasp the essence of data analysis, not only through the tutorial on writing, but more is through practice,
So now it's your turn, use FineBI connected to their data and start the analysis found that the value of the underlying data.