Skip to content

MarcusJul/TrumpTweetsProject

 
 

Repository files navigation

TrumpTweetsPrivate

Econometric Trump Tweet

Next meeting

Venue: LKSLIB Project Room 4-2 Time: 20-Apr-2019 15:00 - 16:00 (1hrs)

Timeline

  1. 4.16: Data collecting
  2. 4.18: Task Distribution
  3. 4.20: Determined data & methodlogy
  4. 4.24: initial result
  5. 4.27: report writting & presentation prepartion

trump tweets source

http://www.trumptwitterarchive.com/archive


result & interpretation

full-sample result & sub-sample result

AAR

  1. positive
  2. negative

CAAR

  1. positive
  2. negative

interpretation

  1. Trumps tweets effect on companyies is more likely a Belated effort (does not change companyies)
    1. postive & negative comparsion
    2. Possibile Reason
  2. Limitation
    1. small sample
    2. may influence industry more
    3. etc

paper 框架

  1. Introduction
  2. Literature Review
  3. Data
    1. Data Source
    2. Data processing
    3. Classfication
  4. Methodology
    1. Event Study
    2. Benchmark
  5. Results:
    1. Benchmark parameter
    2. AAR 以及 CAAR
    3. 可能要除去 earning announcement day 影响再讨论?
  6. Discussion Bias limitation
  7. Conclusion

4.17 进度汇报及安排

讨论内容

  1. 确定只做 individual stock impact
  2. 重新增加并审视了应 drop 的关键词
  3. 丰富了审查的格式(多了一栏 mentioned company name)
  4. 讨论了可以作为 positive negative 判断的依据

成果

  1. 清理出 891 条 tweet
  2. positive/ negative judgement:
    • Positive Normal type
      1. Thanks for investment/ investing
      2. Great meeting with CEO/ manger
      3. Plant/ Job created
    • Negative Normal type
      1. Firms Move to Mexico,
      2. prices are out of control
      3. Threating firms due to Mexico
      4. fake news (CBS ABC CNN NBC)
      5. Bad acts, fines and penalties

接下来进度的安排

  1. 在周四晚上下课前,给 tweet 区分好属性
  2. 周四下课后小聚一会儿,讨论细致分工

当前任务安排

  1. 说明:为了让大家对 Trump tweet 有感觉,每个人有要看的工作量

辛苦夏老师 🌝

  1. 清理指南:
    1. 待清理数据位置:TrumpTweetsProject/code/tweets_data.xlsx (注意,是 excel文件,不是csv
    2. 在第4列(E列)输入数字:1 or -1
      1. 1: postive event type
      2. -1: negative event type
    3. 5列(F列)输入 event tweet 梗概
    4. 无关的事件不做标记
  2. 整理名单 以 excel 的 row index 为序号(不是 A 列
  3. 1-200:ZPH
  4. 201-400: ZZM
  5. 401-600: Lance
  6. 601-800: SL
  7. 801-891: ZLC

positive negative judgement 仅供参考,可能还有新的类型,具体问题具体分析,有疑问可以另行标注后续讨论

About

Econometric Trump Tweet

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 99.7%
  • Python 0.3%