Tagged Location Names in Targeted Social Media Streams

This dataset contains 4500 annotated tweets 1500 tweets from each of three Twitter streams (i.e., Chennai 2015, Louisiana 2016, and Houston 2016 floods). They were tagged using Brat tool recording the start and end character offsets of each mention with a given location category, i.e., inLoc, outLoc, and ambLoc, as mentioned in the LNEx paper.

Applicants to use this dataset are requested to fill in the below form to get the full annotated dataset. After successfully completing the registration form you will be contacted with a link to download the dataset.

This dataset registration information will only be used to (1) check dataset access requests are from genuine researchers and (2) allow occasional follow-up emails after a few years asking you about your experiences using this dataset in your research and, if your experiments were successful, asking if you would be prepared to write a short on-the-record emailed ‘success story’ for us. Any ‘success story’ would only be used for (1) providing examples of successful use for the dataset in advertisement materials and (2) providing evidence of impact in future Kno.e.sis Center proposals / impact assessments relating to use of this dataset.

Kno.e.sis Center is data controller of this dataset.

For any questions and further information contact: Hussein Al-Olimat, hussein@knoesis.org