这篇文章整理自 接地气的陈老师 x 和鲸社区 | 网约车运营分析 数据分析实战活动业务讲解会【接地气的陈老师】的讲解， 更多数据分析动手实践活动欢迎访问>>和鲸社区活动页面
活动介绍
假设你是某打车APP的商业数据分析师，为某大区提供日常数据报表。现在大区领导表示：希望你从日常数据监测中，发现问题和机会点，并做出建议。请你认真研究手头报表数据，给出报告。
示例数据
| 星期 | 时段 | 城市 | 冒泡数 | 呼叫数 | 应答数 | 完单数 | 司机在线 |
| 周一 | 0 | A市 | 29618 | 12616 | 11388 | 11276 | 13700 |
| 周一 | 1 | A市 | 17822 | 7851 | 7025 | 6890 | 9217 |
数据解释
- 冒泡：打车APP中，乘客选择起点、终点，APP提示可选车型&价格，为一次“冒泡”
- 呼叫：乘客看到冒泡信息后，点击“呼叫”按钮，为一次呼叫
- 应答：司机看到乘客呼叫后，接单，为“应答”
- 完单：司机完成订单，乘客付款为“完单”
- 司机在线：每个时段内司机在线人次
首先，我们需要知道企业中的数据分析和学校里学习或者自己练习会的区别：
- 轻过程：企业中数据分析汇报对象是老板，他并不懂得复杂的技术细节，过于专业反而不利于上下级沟通
- 重复枯燥：刚入门的新手大部分处理的都是一些制作报表、临时取数的工作
- 重洞察：需要保持对数据的敏感度，在日常重复的报表中产生洞察
数据集给出的是一个真实环境下使用的报表，这个报表这样设计的原因有：
- 所有的网约车是分城市运营的，每个城市的运营部门都要监控自己城市的运营情况
- 若干个城市会组成一个大区，大区内客观的市场情况可能是差不多的，对于运营情况可以有横向对比
- 时间是和用车是有明显的关系的，比如说周中的早晚高峰，周末出游
After getting familiar with this report, the topic of this event was raised: As a data analyst, what do you see from this report? What does it mean? How to solve this problem?
This is the case in the real work of data analysts. The daily monitoring is usually a simple report, like the one shown in the title, with 3 dimensions and 5 data. You need to find problems in these daily operation data.
Here are 3 problem-solving tips:
- Clarify the criteria: What exactly counts as a problem? This is very different from doing research. When we do scientific research, such as statistics or modeling, they all have some clear standards, such as various test values, such as accuracy. But in actual business, if there are 10,000 orders in this time period, is the 10,000 orders more or less? Can it become a key indicator to evaluate whether a city is functioning well? If not, how should I define the evaluation criteria and justify myself.
- Features of online car-hailing business: The ideal operating state is that drivers and passengers achieve a dynamic balance. If there are too many drivers and not so many passengers, and he cannot receive orders, the driver will definitely be lost; if there are too many passengers and the passengers cannot get a taxi, then the passengers will be lost.
- Reporting skills: Leaders especially like to listen to data, but when you say more than 3 numbers, he will feel that the data is too complicated to understand. So when you want to report to the leader, what should you say? If we only talk about one indicator and one number, what number should we talk about? Try to start from the most important and simplest place, explain a number clearly to everyone, and arouse everyone's interest, then everyone may be interested and can listen to it more.
For more data analysis hands-on activities, please visit >> Hejing community activity page