Brexit Gold Standard

This repository contains the Brexit Twitter Gold Standard produced by the
SSIX Project.



You will need to prepare a few basic things:

  1. Clone this repository git clone ssix-brexit-gold-standard
  2. Move into the cloned repository: cd ssix-brexit-gold-standard/
  3. You need to have both a python interpreter and pip installed
  4. Install required dependecies: pip install -r requirements.txt --upgrade
  5. Create a new Twitter App, then create a
    read-only Access Token, and fill in the missing details at brexit-gs.yml

Rebuild full sample

The published sample
is anonymized to not distribute any personal detail.

To rebuild the full dataset, you need to execute the following script:

python brexit-sample-20160506-annotated.json

This process could take a while, depending on the pause configured at brexit-gs.yml
to obey the Twitter API rate limits.

At the end it willl create a brexit-sample-20160506-annotated-full.json with the full dataset rebuilt,
including both the original tweet data and annotations.


Each Tweet is annotated with the following information:

  1. sentiment: one of "stay", "leave", "undecided", "no sentiment/don't care", "irrelevant", or "" (left blank).
  2. strength: for “stay” or “leave”: an integer between 1 (very weak) and 5 (very strong) expressing the strength of the opinion. For all other tweets, 0.
  3. context: 1 if the interpretation of the tweet depends on external context, such as a linked article or image, 0 otherwise.


Manuela Hürlimann, Keith Cortis, André Freitas, Sergio Fernández, Siegfried Handschuh, and Brian Davis. "A Twitter Sentiment Gold Standard for the Brexit Referendum.".
In Proceedings of SEMANTiCS 2016, Leipzig (Germany), Sep 12-15, 2016.

  title={A Twitter Sentiment Gold Standard for the Brexit Referendum.},
  author={H{\"u}rlimann, Manuela and Cortis, Keith and Freitas, Andr{\'e} and Fern{\'a}ndez, Sergio and Handschuh, Siegfried and Davis, Brian},
  booktitle={12th International Conference on Semantic Systems Proceedings},

Brian Davis, Keith Cortis, Laurentiu Vasiliu, Adamantios Koumpis, Ross McDermott, and Siegfried Handschuh. "Social sentiment indices powered by X-scores.".
In 2nd International Conference on Big Data, Small Data, Linked Data and Open Data (ALLDATA 2016).
More information,
Google Scholar,

  title={Social Sentiment Indices Powered by X-Scores},
  author={Davis, Brian and Cortis, Keith and Vasiliu, Laurentiu and Koumpis, Adamantios and McDermott, Ross and Handschuh, Siegfried},
  booktitle={ALLDATA 2016, The Second International Conference on Big Data, Small Data, Linked Data and Open Data},



The software is available under the business-friendly license
Apache License, Version 2.0.
Therefore, you are completely free to use the software for any purpose,
to distribute it, to modify it, and to distribute modified versions of the software,
including closed-source, under the terms of the license, without concern for royalties.


All of the data from Twitter (tweets, creation dates, tweet ids) are covered by Twitter’s Terms of Service.

The annotations are licensed under the Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) license.

Please cite [Hürlimann et al. 2016] in all publications using this dataset (see References).


If you have any questions about this repository, please contact Sergio Fernández.
For questions about the data set and annotations, get in touch with Manuela Hürlimann.
For any other general question of the SSIX Project, you are welcome to contact us.


This work is in part funded by the SSIX Horizon 2020 project
(grant agreement No 645425) and Science Foundation Ireland (SFI) (grant number SFI/12/RC/2289).