Data Scientist 的工作分布、Team Capability 與 Data Preparation Reality。
理解 Data Scientist 不只建模。
原書引用 O’Neil and Schutt：Data Shaping 可能占到 90% 時間；O’Reilly 2017 Survey 近 800 participants、69 countries。
| 編號 | Task | Major Involvement |
|---|---|---|
| 1 | Basic exploratory data analysis | 67% |
| 2 | Conduct data analysis to answer research questions | 61% |
| 3 | Communicate findings to business decision-makers | 58% |
| 4 | Data cleaning | 53% |
| 5 | Develop prototype model | 49% |
| 6 | Create visualizations | 47% |
| 7 | Identify business problems that can be solved with analytics | 44% |
| 8 | Feature extraction | 42% |
| 9 | Organize and guide team projects | 40% |
| 10 | Implement models/algorithms into production | 38% |
| 11 | Collaborate on code projects（read/edit others’ code, using the git repository） | 38% |
| 12 | Teach/train others | 31% |
| 13 | Communicate with people outside your company | 30% |
| 14 | ETL（extract, transform, load） | 29% |
| 15 | Plan large software projects or data systems | 28% |
| 16 | Develop dashboards | 28% |
| 17 | Set up/maintain data platforms | 24% |
| 18 | Develop data analytics software | 21% |
| 19 | Develop products that depend on real-time data analytics | 18% |
| 20 | Use dashboards and spreadsheets（made by others）to make decisions | 15% |
| 21 | Develop hardware（or work on software projects that require expert knowledge of hardware） | 4% |
Excel 排序 Task、計算 pp Gap；AI 解釋 Skill Coverage。
67% − 4% = 63 pp；58% − 49% = 9 pp。
Data Science 是 Team Sport；Business Communication 與 Data Preparation 都是核心工作。
「找 Team Capability Gaps；不要把 Major Involvement% 解釋成工時%。」
題目： 只聘 Modeling 強者會漏什麼？
參考解答： Business Communication、Problem Identification、Data Preparation、Deployment 等。