How to provide insight for the enterprise by Power BI

    在如何修炼成一个合格的商业分析师中我们提到一个好的商业分析师是要通过工具快速分析出事件的根源，为企业最终的决策提供有价值的洞察力。今天我们来和大家一起探究如何通过Power BI为企业提供有价值的洞察力。下面是一个有关人事部门数据的仪表盘分析，希望对大家有所启发：
            此前领导迷恋的全民热剧《人民的名义》，经常拿剧中李达康书记在懒政干部培训班上的演讲来教育我们企业中万万不能有“懒政”现象。那如何去判定企业中谁有”懒政”现象呢？如何去判断企业中 “懒政”现象是否已经有恶化的趋势呢？我们用Power BI做了一个仪表板来告诉他！（本文所涉及的任何数据均为测试数据）
            ![](https://s1.51cto.com/images/blog/201908/06/419eb08602be978fcd42b5e53d0601ca.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
            最近H公司第三季度的完成额不是很漂亮，大boss非常不高兴，布置任务给了公司分析师小W要求找出问题的症结。
            小W:老板，您交代的任务完成了，通过分析，我发现了公司第三季度完成额不理想的原因，我通过演示来告诉您问题出在哪。
            大Boss：OK。
            小W:当我们点开人事部报告，我们从公司人员组织架构的主要信息中发现公司管理岗(Manager)和员工岗(staff)的比例相近（下图1.0），说明可能存在管理岗人员过多的问题。
            ![](https://s1.51cto.com/images/blog/201908/06/25ae4348b4619f85bf0b9013ac428a78.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
            大Boss：嗯，说的对，看一下具体哪个部门人员比例失调。
            小W:老板，Jenny部门确实管理岗人员比率有点高（下图2.0）。
            大Boss：看一下Jenny部门第三季度的人员绩效。
            小W:好的老板，通过仪表盘右边的人员绩效，发现Jenny团队中的老员工(工龄超过一年)的业绩非常不好（下图2.1），而且人数超过70%（下图2.2）。![](https://s1.51cto.com/images/blog/201908/06/74e7593f1378b4c82be814fd212cdc11.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)

            ![](https://s1.51cto.com/images/blog/201908/06/74e7593f1378b4c82be814fd212cdc11.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
            大boss：看来Jenny团队确实存在问题。下周例会上需要和Jenny谈谈了。
            所有这些过程，我们仅设计了一个仪表板就发现了员工中“懒政”现象的存在，对未来的员工管理提供了有效的数据支持。仪表盘是一种可以有效分享洞察力的工具。通常情况下，仪表盘没有很大程度的被人挖掘出它的价值，因为很少有人把它放在数据可视化中。然而当你用过仪表盘后，你就可以真正体会到power BI的价值所在。
            有商业价值的仪表板设计既是一门艺术也是一门科学。仪表盘具有灵活性，多变性，丰富性的特点，可以按照分析师的想法呈现出不一样的展示效果。但如果你传递的信息是杂乱无章的，分散的，错误的，那么做出来的东西也就是没有任何意义。任何呈现数据的人都应该努力让他们的目标受众容易理解所做的仪表盘。任何一个简单的设计更改，都有可能会在内容和清晰度方面产生巨大的差异。
            如果你没有时间去学习制作仪表盘，这里有一些小贴士希望可以帮助到你：
            1.  考虑你的用户：

• you must first know the dashboard is used to do.
• What information the user needs.
• How to focus the information required to perform the next step.
• personal preferences or cultural background will not affect the overall design.
• What is the meaning color.
• which icons make it easier for users to accept.

2. Best use dashboard design:
   • Story line: good design should tell a story with data, the related items together, do not put too much information to the target user useless, messy visual controls affect users understand the business.
   • Keep it simple dashboard: page only the best 3-5 values, charts or tables. Avoid placing too much information on the dashboard. A good dashboard should one shows all relevant information without having to scroll or click.
   • recognize the user's reading habits, each location has its own meaning and different levels of importance. Do not randomly placed chart, if you need more tables, place them on the bottom of the dashboard.
   • Uniform Color:, dimensional color map and sequence charts dimension values should be consistent with the axis on the chart.
   • Font unity: page Filters font, title font, the size of the font, text format should be unified.
   • Select the appropriate data visualization controls: Do not use easy to understand charts that are not commonly used, such as 3d charts. The human brain is difficult to explain the shape of a circular, pie, donut FIGS, scale, and other types of circle graph it may be more beautiful than, but they are not the best data visualization Fig.
   • Be sure to properly display the data: not more than 3 or 4 digit numeric display. After the decimal point 1 or 2 digits to display as thousands or million or 3.4 million instead of 3.4 million.
   • Do not mix with precision and time. Ensure that the time frame is fully understood. For example a chart, show the data for a particular month, in which case the screening time, your table will be filtered out.
   • Do not put large and small measurements are mixed together, such as in a line or bar graph, a measure is the number of millions, while the other measure is thousands. In such a large scale, it is difficult to see the difference at thousands of metric.
   • Do not use unnecessary data labels to influence the chart. The length of the bar graph is generally easier to understand than the digital.
   • align elements: If you want to do the dashboard looked very professional, then align elements from the start. The border can be opened in the production of the dashboard, it puts a good visual controls each row or column may be in their element by alignment. Of course, if you do not like borders, in've created a dashboard button can turn off the border.
   • Avoid abbreviations does not make sense: unless the target users know the meaning of an abbreviation, it is recommended that you use the full name or display name, it is easier to understand the meaning of the target user, rather than guess.
   • Display changes: the most intuitive to the user data, do not let the user do the math.
   • Design a dashboard, instead of reporting: Remember that not all the details in the table are applied to the instrument panel.
   Today's share on here. For more articles in the business dashboard, click Microsoft Power BI classic case of inspection.