Analytic Approach Based on the Question Type基于问题类型的分析方法

When choosing an analytic approach for a problem, the type of question you’re trying to answer greatly influences the methodology. Here are five common types of questions and corresponding analytic approaches:
在选择针对问题的分析方法时，你试图回答的问题类型会极大地影响方法论。以下是五种常见的问题类型及其对应的分析方法：

1. Descriptive Questions: “What is the current status?” 1. 描述性问题：“当前状态如何？”

Approach: Descriptive Analytics
方法：描述性分析

Question: "What is the current status of our sales?"
问题："我们目前的销售状态如何？"

Techniques:  技术：

• Data aggregation: Combining data from various sources into a unified view.
  数据聚合：将来自不同来源的数据整合到一个统一的视图中。
• Data mining: Extracting useful information from large datasets.
  数据挖掘：从大型数据集中提取有用信息。
• Data visualization: Using visual tools to present data in an easily understandable format.
  数据可视化：使用视觉工具以易于理解的方式呈现数据。

Examples:  示例：

• Summarizing sales data  总结销售数据
• Creating dashboards  创建仪表盘
• Generating reports  生成报告

2. Diagnostic Questions: “Why did it happen?” 2. 诊断问题：“为什么会发生这种情况？”

Approach: Diagnostic Analytics
方法：诊断分析

Question: "Why did our sales decline in the last quarter?"
问题：“我们上一季度的销售额为什么会下降？”

Techniques:  技术：

• Drill-down: Exploring detailed data to find underlying causes.
  下钻：探索详细数据以发现潜在原因。
• Data discovery: Identifying patterns and relationships in data.
  数据发现：识别数据中的模式和关系。
• Correlation analysis: Assessing the relationship between different variables.
  相关性分析：评估不同变量之间的关系。

Examples:  示例：

• Identifying root causes of sales decline
  识别销售额下降的根本原因
• Analyzing customer complaints
  分析客户投诉
• Understanding failure points in a process
  理解流程中的失败点

3. Predictive Questions: “What is likely to happen?” 3. 预测性问题：“可能会发生什么？”

Approach: Predictive Analytics
方法：预测分析

Question: "What is our sales forecast for the next year?"
问题："我们下一年度的销售预测是多少？"

Techniques:  技术：

• Regression analysis: Predicting outcomes based on relationships between variables.
  回归分析：根据变量之间的关系预测结果。
• Time series forecasting: Predicting future values based on past trends.
  时间序列预测：根据过去趋势预测未来值。
• Machine learning models: Using algorithms to predict future outcomes based on historical data.
  机器学习模型：使用算法根据历史数据预测未来结果。

Examples:  示例：

• Forecasting sales  预测销售额
• Predicting customer churn
  预测客户流失
• Estimating future demand  估计未来需求

4. Prescriptive Questions: “What should we do?” 4. 规范性问题：“我们应该做什么？”

Approach: Prescriptive Analytics
方法：规范性分析

Question: "What should we do to increase website traffic?"
问题："我们应该如何增加网站流量？"

Techniques:  技巧：

• Optimization models: Finding the best solution from a set of alternatives.
  优化模型：从一组备选方案中找到最佳解决方案。
• Simulation: Modeling scenarios to predict outcomes.
  模拟：建模场景以预测结果。
• Decision analysis: Evaluating and comparing different decisions.
  决策分析：评估和比较不同的决策。

Examples:  示例：

• Recommending inventory levels
  推荐库存水平
• Optimizing marketing campaigns
  优化营销活动
• Determining pricing strategies
  制定定价策略

5. Classification Questions: “Which category does this belong to?” 5. 分类问题：“这个属于哪个类别？”

Approach: Classification (Supervised Learning)
方法：分类（监督学习）

Question: "Which category does this data point belong to?"
问题：“这个数据点属于哪个类别？”

Techniques:  技术：

• Logistic regression: Predicting the probability of a categorical outcome.
  逻辑回归：预测分类结果的概率。
• Decision trees: Splitting data into branches to classify it.
  决策树：将数据分割成分支进行分类。
• Support vector machines: Finding the best boundary to separate categories.
  支持向量机：寻找最佳边界以分离类别。
• Neural networks: Using interconnected nodes to classify data.
  神经网络：使用相互连接的节点对数据进行分类。

Examples:  示例：

• Email spam detection  电子邮件垃圾邮件检测
• Image classification  图像分类
• Disease diagnosis  疾病诊断

Understanding these different types of questions and the corresponding analytic approaches can help you unlock your data's true potential.
理解这些不同类型的问题以及相应的分析方法，可以帮助你解锁数据的真正潜力。