TAGGS: Grouping Tweets to Improve Global Geoparsing for Disaster Response

Abstract

Timely and accurate information about ongoing events are crucial for relief organizations seeking to effectively respond to disasters. Recently, social media platforms, especially Twitter, have gained traction as a novel source of information on disaster events. Unfortunately, geographical information is rarely attached to tweets, which hinders the use of Twitter for geographical applications. As a solution, geoparsing algorithms extract and can locate geographical locations referenced in a tweet's text. This paper describes TAGGS, a new algorithm that enhances location disambiguation by employing both metadata and the contextual spatial information of groups of tweets referencing the same location regarding a specific disaster type. Validation demonstrated that TAGGS approximately attains a recall of 0.82 and precision of 0.91. Without lowering precision, this roughly doubles the number of correctly found administrative subdivisions and cities, towns, and villages as compared to individual geoparsing. We applied TAGGS to 55.1 million flood-related tweets in 12 languages, collected over 3 years. We found 19.2 million tweets mentioning one or more flood locations, which can be towns (11.2 million), administrative subdivisions (5.1 million), or countries (4.6 million). In the future, TAGGS could form the basis for a global event detection system.

A pragmatic guide to geoparsing evaluation

“…'', the geoparsing pipeline is (1) Toponym extraction [Springfield, Waldo County Jail, Maine], this step is called Geotagging and is a special case of NER; and (2) Disambiguating and linking toponyms to geographic coordinates [(45.39, -68.13), (44.42, -69.01), (45.50, -69.24)], this step is called Toponym Resolution (also Geocoding). Geoparsing is an essential constituent of many Geographic Information Retrieval (GIR), Extraction (GIE) and Analysis (GIA) tasks such as determining a document's geographic scope (Steinberger et al 2013), Twitter-based disaster response (de Bruijn et al 2018) and mapping (Avvenuti et al 2018), spatio-temporal analysis of tropical research literature (Palmblad and Torvik 2017), business news analysis (Abdelkader et al 2015), disease detection and monitoring (Allen et al 2017) as well as analysis of historical events such as the Irish potato famine (Tateosian et al 2017). Geoparsing can be evaluated in a highly rigorous manner, enabling a robust comparison of state-of-theart (SOTA) methods.…”

A Citizen-Sensing-Based Digital Service for the Analysis of On-Site Post-Earthquake Messages

“…Many studies have proposed solutions to the problem of determining the origin of tweets (e.g., [4,15,16]). Their approaches range from machine learning, statistics, probability, natural language processing to geographical information systems and surveying.…”

“…Unfortunately, "identifying the location from where the information is being disseminated is still a non-trivial issue", (Laylavi et al [3], p. 2). de Bruijn et al [4] have published an interesting paper about the geoparsing of groups of related tweets. Their method improves previous proposals devoted to the identification of the location of tweets; nevertheless, authors state that about 10% of the predicted locations are incorrect.…”

Mine Water Problems and Solutions in China

“…Interdisciplinary and multidisciplinary research must be promoted. For example, metadata and the contextual spatial information gathered and analyzed by geographical information techniques can be used for tracing disaster events (de Bruijn et al 2018). This technique can be adopted by relief organizations to respond to mine water disasters.…”