📈

数据科学家面试

统计 · SQL · ML · 实验设计 · case

6 科目 DS / Analyst / Quant 面试Advanced

原创复习卡。统计与 ML 公式请以教科书最新版为准。

考试结构 · 点击右侧按钮生成该科目卡组

#01

统计基础

假设检验 · 分布 · 抽样 · 置信区间

p-value interpretationType I vs Type II errorconfidence intervalscentral limit theoremBayesian vs frequentist+2

#02

SQL 提问

join · window · CTE · 优化

join types and when to use eachwindow functions (row_number, rank, lag/lead)CTEs vs subqueriesGROUP BY with HAVINGself-joins for hierarchies+2

#03

机器学习基础

回归 · 分类 · 评估 · 过拟合

bias-variance trade-offregularization (L1 / L2)logistic regression vs lineardecision trees and ensembles (RF, GBM)precision / recall / F1+2

#04

实验设计 (A/B)

样本量 · 显著性 · 二级指标 · 长期效应

minimum detectable effectsample size calculationnovelty and primacy effectsstratified samplingswitchback experiments+2

#05

Case 题

metric drop · 推荐设计 · 实验解读

DAU dropped 5% — diagnosedesign a recommendation systemfraud detection feature designinterpreting confusing A/B resultsmetric for a search ranking change+1

#06

行为题

影响力 · 不确定性 · 协作

STAR frameworkpushed back on a request you disagreed withambiguous data you had to make decisions onpresenting findings to non-technical audiencebiggest analytical mistake