Skip to content

Fitzhugh et al – 2016 – Spatio-temporal filtering techniques for the detection of disaster-related communication

Summary: Quantitative study that explores a protocol for spatio-temporal filtering of user-generated natural disaster content (referred to here as informal communication) on Twitter.

Synthesis: Temporal filtering is operationalized as tweet timestamps and successive periods of 24-hr clock time. Some methodological decisions in this paper give me pause in citing it in my own work.

Foundational concepts in this study: sociotemporality, geolocation,

Agreement in related work: N/A

Contested areas: Acknowledges the difficulty of collecting sensemaking attributes, like temporality, in informal user-generated content during a disaster. Processing high signal-to-noise ratio communications is a long-standing problem.

Gaps/Limits in this study: Much of the related work in rumoring/communication is quite old — conducted upwards of 20-60 years ago. None of it conceived in a hyper-networked online society. This is not my area of expertise and perhaps this topic has not been recently addressed which forced the authors to situate this study in long-established theories/frameworks. However, that the age of the related work was not raised as a potential limitation is concerning. Likewise, the paper only evaluated geolocated tweets which are < 1 percent of all posts. That should have been disclosed and grappled with in the discussion. I question how relevant they are to current behavior -- either offline or online -- when ICT devices are vastly different now than in the 1940s. With such a negative connotation to the word "rumoring" I'm surprised that the authors are not simply using "informal communication" or suggesting a more nuanced term. By this definition, digital humanitarian work would be described as rumoring. This seems really problematic. Connections to my work: sociotemporality, social media, crisis informatics

Annotated paper:

Fitzhugh et al. – 2016 – Spatio-temporal filtering techniques for the detection of disaster-related communication-annotated

Published inannotationstemporality