Choose Analysis Scenarios

Scenario 1: Cross-sectional sales analysis

A sales director reviews revenue, gross profit, and collections by region × product line × month each week. Create sales_perf_view:

display_name: 销售业绩分析视图
physical_table: dws_sales_perf_monthly
dimensions:
  - name: region
    description: 销售大区
  - name: product_line
    description: 产品线
  - name: month
    description: 月份（YYYY-MM）
metrics:
  - name: deal_amount
    description: 成交金额
  - name: gross_profit
    description: 毛利
  - name: collect_rate
    description: 回款率
llm_description: |
  这是一个预聚合的销售业绩分析视图，按大区、产品线、月份切分，
  用于高性能地分析销售成交金额、毛利和回款率。
  适合销售总监周会、季度复盘等横切分析场景。

A question such as "Gross profit for electronics in East China in the first half of the year" directly matches this view.

Scenario 2: Project workforce productivity review

Project reviews examine cost, person-days, and risk counts by project × month. Create a view with Project ID and Summary Month as dimensions, and Monthly Total Cost, Monthly Person-Days, and Risk Count as measures. Ask the position during the review meeting and get answers without waiting for the data team to extract data.

Scenario 3: Cross-sectional customer health analysis

Customer success teams track activity, renewal rates, and NPS by customer × quarter. Use the customer object (Customer) as a dimension and link the team's preaggregated wide table to create customer_health_view. An account manager can @ a customer in Chat and ask about its health trend; AI queries the view directly.