Reputation: 436
For example, getting the information that there were 56000 tweets in Los Angeles, CA on May 30th, 2015 at 5:00 PM. Ideally, I would like to access tweets as far back in time as possible. The main issue with this seems to be getting tweets from the past - Twitter's REST API only has tweets for about the past 7 days.
The only way that I am thinking of is manually counting the tweets for a particular location and time from Twitter's website directly, but surely there must be a better way.
Is there some simple way to accomplish this that I'm not thinking of, or any libraries or frameworks that optimize this task?
As of now, I'm trying to use R, but I'm not sure if other languages or frameworks would be better suited for this.
Upvotes: 0
Views: 390
Reputation: 4058
You will never achieve your goal, even if you pay for Gnip. The reason is less than one to two percent of twitter accounts reveal their location.
If that is not a problem for you, and you don't want to pay for Gnip, you can use statuses/filter, which is a streaming endpoint, along with the locations
parameter. Then you would simply count the tweets as they come. However, you would not have access to historical information.
Gnip would also give you the historical tweets you desire, and Gnip has some intelligent algorithms that will increase the number of geo-locatable tweets to around 20%, more or less.
Upvotes: 1