When and Where Did it Happen? An Encoder-Decoder Model to Identify Scenario Context