#1

Hello,

Thank you for approving the developer account access! In preparing for the summer data journalism project, we are realizing that the free account will probably not cover everything we want to do, and so we have looked into the paid accounts. Unfortunately, we are on a tight budget, and the monthly payments are pretty expensive for us, being three college students with only limited funding. Is there any type of educational discount that could be provided?

Thank you!

#2

Thanks for reaching out, currently we don’t have such a discount but feel free to submit this an idea on our feedback channel.

Can you give us a bit more context about what you are trying to accomplish here, we might be able to help out.

Best,

Jessica

2 Likes
#3

What’s the data you want to retrieve? There’s more than likely ways of getting what you need with a combination of REST APIs and Sandbox Premium to make a proof of concept.

1 Like
#4

We are trying to analyze the partisanship of different news organizations by estimating the partisanship of each of their followers by seeing which political accounts they follow. We are a little unsure of how queries for followers impact our monthly limit since we are not querying actual tweets, but some of these organizations have tens of thousands of followers, and then each of their followers has a lot of followers, so these numbers add up quickly. Do you have any ideas on the best way to do this or info on how queries like this will impact our monthly API limit?

#5

See post below. Do you have any ideas on the best way to do this or info on how queries like this will impact our monthly API limit?

#6

Queries to the standard API friends and followers endpoints have no impact at all on the monthly quota for the premium search APIs.

#7

I would start with an existing dataset - like this one GitHub - pablobarbera/twitter_ideology: Estimating Ideological Positions with Twitter Data

To make your own dataset will involve using these endpoints:

Each endpoint has it’s own separate limit of 30 calls every 15 minutes. (I’m assuming that you will be using both the user context, 15 calls - with an access token, and app only auth, another 15 calls with a bearer token OAuth 2.0 App-Only (Bearer Token) | Docs | Twitter Developer Platform )

You should always set the count to the maximum: /ids endpoints can get 5000 ids in 1 single call, and you can get the full user objects with GET users/lookup | Docs | Twitter Developer Platform. /list endpoints get 200 in 1 call and include the full user object. In many cases, it’s possible to get all friends and followers of an account with a single call to /ids endpoints for friends and followers.

Friends and followers are returned in roughly reverse chronological order, so the newest followers / friends will be returned first.

Given that /users/lookup gives you friend and follower counts, you can decide and optimize your retrieval to get the most out of your calls.

You could also give a budget of calls per user, spending a maximum of 1 call per user, saving the cursor and continuing collection later - this way you will cover a wider number of accounts, and won’t get stuck on extremely popular ones for long.

For extremely popular accounts, you can use GET friendships/show | Docs | Twitter Developer Platform to check if a user is following another without having to retrieve all of them. You can call this 195 times every 15 minutes.

All of this is in the Standard REST API, which is not paid and not connected to the Premium API. There is no paid version of the endpoints that return friends and followers, so you need to budget a good amount of time for data collection if you have a lot of users and a lot of friends and followers to crawl.

#8

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.