The dataset contains the well balanced disaster and none disaster tweets selected from 2011, 2012, 2013, 2014, 2015, 2017 and 2018.
The predicted label is shown in predict column