API terms of service


The developer rules of the road include an commandment that one not attempt or encourage others to sell, rent, lease, sublicense, redistribute, or syndicate access to the Twitter API or Twitter Content to any third party without prior written approval from Twitter. To whom should I write with questions about whether I need to obtain written approval from twitter for a website I’m building as part of a research project.


Exceptions are not typically granted for this rule. For research projects, you cannot redistribute Twitter data. If you want to provide evidence of your research, you can provide tweet IDs, user IDs, and screen names but you cannot provide actual tweet or user objects nor the text property of tweets. You must require those who wish to reproduce your research to use the API to re-vivify the IDs into proper objects.


Thanks for your help Taylor. My question was vague, I wasn’t certain who might be listening, please allow me to be specific.

My research group is building an instrument to remotely sense population-level sentiment that we call the Hedonometer. We recently published our analysis covering several years of the gardenhose feed, a study indicating that happiness has dropped over the past few years due to an increase in negative words. Here is the publication:

Along with many other research endeavors related to measuring happiness (e.g. network effects), we are now interested in building a website that would provide visitors with the output of our instrument in near real-time, using graphics similar to the top panel of figure 3 from the paper, an image you can view by clicking here:


The website would not allow visitors to see twitter messages or user information, it would simply present a simple timeseries of happiness as measured by our algorithm, along with a list of the words responsible for shifts in happiness from day to day. The information presented is aggregate, involving millions of users in any single calculation. As such, I’m confident that the website won’t violate the API terms of use.

The gardenhose feed has outgrown UVM’s ability to provide the required computational resources, and we need to move our analyses off of campus hardware in order to continue making progress on the project. The website construction has been funded in part by the MITRE corporation, with whom we are collaborating on a formal partnership as researchers at the University of Vermont. My question regarding the TOS relates to whether we are allowed to store the JSON files on an Amazon cloud server paid for by MITRE. This server would feed the website, and be available for our own research purposes. The raw JSON would be available only to members of our research team and as such are not being redistributed.

Thank you for your time, I look forward to hearing your thoughts.


@episod Taylor, is this the correct venue in which to interact regarding the issue of appropriate use of the gardenhose? Thanks for your time.


You’ll probably want to follow up with our API policy team for any solid policy determination on this. https://support.twitter.com/forms/general

I’ll offer you my opinion and own interpretation of rules in this context though:
If the scope is limited to your research purposes and application use case at UVM, the storage of JSON on a secured Amazon Cloud Server with your service as its only customer is an allowed usage – your cloud server is acting as an extension of your own network’s storage capabilities.


Thanks for your thoughts Taylor!