joe5saia / joe5saia/eventstudiesdata

Speeches are messy

Open
#3 2 comments 0 reactions 1 assignee Claimed by @miguel-acosta View on GitHub
Dominant language
Python
Stars
2
Forks
1
PR merge metrics
No merged PRs in 30d

Description

### The issues
Bloomberg's database of FOMC speeches still needs some cleaning. In many cases the same speech/date/time is listed multiple times--we can safely drop duplicates here since the date/time is the same.

Of more concern are when a single person gives multiple speeches on one day--while this is possible in practice, some spot-checking suggests that this may be a mistake. Total this gives us 708 duplicates of 5820 speeches.

If we only keep speeches categorized as "Central Banks", the duplicate ratio goes to 255/5240. This eliminates many of the 708 cases in which the "duplicates" where within a few (1-10) minutes of one another.

### To ask Bloomberg

1. Where is the data from?
2. Why do the non-categorized speeches have a slightly different timestamp than the categorized ones? How are speeches categorized?
3. Bring up a few cases where it is unlikely---based on external evidence---that someone gave two speeches on one day (Moskow at Ball State is a good one).

Contributor guide

No contributing guide indexed for this repository

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.