Text Mining Methodology for Building Water Leak Maps from Tweets

Abstract

This article presents a text mining methodology that is used in the framework of a project called Fugometría. The objective of the proposed methodology is the construction of maps with markers that indicate the location of water leaks. To achieve this purpose, the initial step is the collection of tweets issued by citizens to report the existence of leaks in drinking water distribution networks. Once these tweets are collected, they are parsed for an address in the body of the message. If an address is detected, it is converted into a GPS coordinate, which is in turn used to build a map with markers indicating the location of water leaks. In order to show the applicability of the methodology, some preliminary results on the construction of a leak map of Mexico City are shown.