Add new values to the tweet object: direct_favorite_count and direct_retweet_count

retweet
api

#1

I want to make the case that adding the values direct_favorite_count and direct_retweet_count to the tweet object would be useful not just to me, but also to other users of the API, regular twitter users, and to twitter itself.

But first, here’s a little context: I am making an app that finds a user’s best tweets by measuring tweet “quality” using the values in favorite_count and retweet_count. My simple equation is:

quality = tweet.favorite_count + tweet.retweet_count

But there’s one big issue with this: retweets. In a retweet, favorite_count and retweet_count apply to the original tweet, not the retweet. They count favorites/retweets on the original tweet, plus favorites/retweets on any retweet of it. That means that favorite_count and retweet_count for retweets are vastly inflated compared to non-retweets, so a straight comparison is impossible.

There’s also a second, smaller issue: Let’s say I make a tweet that gets only 1 favorite from my followers. But a user with many followers retweets it, and his retweet gets 1000 favorites. My tweet now has a favorite_count of 1001, even though it was barely relevant to my own followers.

In summary, the retweet system makes it hard to figure out whether a tweet is actually relevant to followers of the user, or if the favorites and retweets came from somewhere else.

That’s why I propose adding two new values to the tweet object: direct_favorite_count and direct_retweet_count. These values should measure only favorite and retweet clicks that happened directly on this tweet. Not clicks on a retweet of it. Not clicks on the original tweet of a retweet. Just clicks on this tweet directly.

This gives us some new and interesting data. Now we can see clearly whether a tweet/retweet is relevant to the followers it was tweeted to. And this data is not just useful for the API. It’s also useful to users trying to make their tweets more relevant, and useful to twitter itself, helping to analyze follower demographics for better advertising, recommendations, etc.

I’m assuming this would require a change to the twitter database, and not just to the API. But for the reasons stated above, I believe it is worth it.