Bogota city traffic accidents - Social media datasets
Description
The main data set is composed of 4,538,305 Twitter post in Spanish language related to road accidents gathered during the period between October 2018 and July 2019 in Bogotá city, Colombia. The data was collected using the Twitter Search and Stream APIs, using a combination of traffic accident keywords and the information made available by users such as traffic police, transportation companies and local government agencies from Bogotá city. The collected raw data was processed in order to remove duplicate entries and generate a dataset free of non-printable characters, stop worlds and non-relevant information, using techniques such as lemmatization and stemming. Finally, a sample of 9640 registers of the dataset was labelled to discriminate if the record is relevant to road accident information. The proposed used of the dataset is academic research of traffic accident detection and real time traffic modelling.