Sweeping Away the Noise: An Interactive Subtractive System for Disaster Information Management on Social Media

Abstract

While social media serves as an essential social sensor, its data stream is contaminated with massive noise, causing severe information overload. To address this challenge, this research proposes an interactive interface based on a novel “subtractive strategy” that actively eliminates noise. The core feature is a “stepwise exclusion” mechanism, which enables users to iteratively filter out irrelevant clusters of posts grouped by the content of their attached images. To verify the system’s efficacy, we conducted a user study using a dataset of tweets from the July 2020 heavy rainfall event in Japan. Compared to a simple baseline system (text and image presentation), the proposed system collected more damage-confirming posts, with the difference emerging in the final third of a 30-minute session.

Information