A Dataset of UN Agencies' Public Communication about Climate Change on Twitter

The present dataset contains the Twitter communication of eight international organizations (IOs) in different policy areas that are known to be central in communicating about climate change. The IOs are comparable in their communication, all being parts of the United Nations (UN). The IOs under consideration are:

• Food and Agriculture Organization (FAO),
• Office for the Coordination of Humanitarian Affairs (UNOCHA),
• UN Development Programme (UNDP),
• UN Office for Disaster Risk Reduction (UNDRR),
• UN Environmental Program (UNEP),
• UN International Children’s Emergency Fund (UNICEF),
• UN High Commissioner for Refugees (UNHCR),
• World Health Organization (WHO).

The tweets were downloaded and parsed via the Twitter Academic Research API (link). In total, the dataset contains 222,191 tweet IDs of the tweets posted by the above 8 UN organizations from their official accounts. This number represents the total number of tweets posted by these selected UN organizations since the beginning of their tweeting history until the end of 2019. The dataset is compliant with the privacy policy, developer agreement, and guidelines for content redistribution of Twitter and the FAIR principles (Findability, Accessibility, Interoperability, and Reusability) principles for scientific data management.

The dataset consists of two parts:

• Unlabeled tweet IDs of the considered IOs (8 txt-files),
• Labeled dataset of tweet IDs with labels indicating whether tweets are about climate change or not (1 csv-file).

Unlabeled tweet IDs

The corresponding 8 txt-files contain tweet IDs of the corresponding tweets posted by the UN organizations. The files are summarised in Table 1 below. 

The dataset contains only tweet IDs to ensure compliance with the terms and conditions mentioned in the privacy policy, developer agreement, and guidelines for content redistribution of Twitter. The tweet IDs need to be hydrated to be used. For hydrating the present dataset, the Hydrator application (link) may be used; see a step-by-step tutorial on how to use Hydrator (link).

Labeled dataset related to climate change

This is a subset of the entire dataset described above. Namely, 5,750 tweets are randomly selected from the entire dataset and labeled manually as either "climate change-related" or "not climate change-related". The dataset is available in the file dataset_UN_climate_change_labeled.csv and is summarised in Table 2 below.