#1

I am wondering if there has been a change to retweet_count logic of the public_metrics (possibly with the release of the new retweet endpoint)?

Could someone confirm the retweet_count of a retweet? The expectation of this value is the number of times a retweet was retweeted. I am now seeing this value matching the original tweets retweet_count.

For example:
User A posts Tweet 1
10 users retweet Tweet 1
User B retweets Tweet 1
User C retweets User B’s retweet

What is the retweet_count of User B’s retweet?

  • Previously, I would expect 1 (from User C retweeting)
  • Now, I believe it will be 12 (10 retweets + User B retweet + User C retweet)

Hopefully this makes sense.

#2

How do you mean changed? Do you have an example json? I think it was always the same - the retweet_count for a tweet, is always the global retweet count, never a “retweet of a retweet”

So User C’s retweet should read 12.

1 Like
#3

Apologies, I should have provided an actual example.

Referring to the result from this api call:
https://api.twitter.com/2/tweets/1506673101072998407?tweet.fields=id,public_metrics,text&expansions=referenced_tweets.id

Result:

{
	"data": {
		"referenced_tweets": [
			{
				"type": "retweeted",
				"id": "1505977175488483339"
			}
		],
		"id": "1506673101072998407",
		"text": "RT @elonmusk: Excited to hand over the first production cars made by Giga Berlin-Brandenburg tomorrow!",
		"public_metrics": {
			"retweet_count": 11604,
			"reply_count": 0,
			"like_count": 0,
			"quote_count": 0
		}
	},
	"includes": {
		"tweets": [
			{
				"id": "1505977175488483339",
				"text": "Excited to hand over the first production cars made by Giga Berlin-Brandenburg tomorrow!",
				"public_metrics": {
					"retweet_count": 11604,
					"reply_count": 10450,
					"like_count": 201853,
					"quote_count": 876
				}
			}
		]
	}
}

In the response, you’ll note that the retweet_count is the same:
data.public_metrics.retweet_count: 11604
includes.tweets[0].public_metrcis.retweet_count: 11604

I believe (although I can no longer prove it with an actual example), that the value of data.public_metrics.retweet_count should refer to retweets of the retweet, and would be 0 in this case.

Instead, the value matches the retweet_count of the original tweet.

Hopefully this example clears up any ambiguity.

For additional context, I am interested in determining the popularity / engagement of a tweet. In this case, there is a false positive if using the retweet_count only as a metric.

#4

Oh thanks, yeah that makes sense - unfortunately this is not the case, the retweet count is always the global rt count of the original tweet only.

1 Like
#5

Do you know if that’s always been the case? I believe it has not always been the case (within the past few months even). Also, I feel there’s some inconsistency as well. I would expect the like_count to behave the same. This implies you could like a retweet, but not retweet a retweet.

#6

Yeah, as far as could tell this was always the way. I’d love it if it was the other way though.

1 Like