From Social Media to Ballot Box:Leveraging Location-Aware Sentiment Analysis for Election Predictions

Predicting election outcomes is a crucial undertaking,and various methods are employed for this purpose,such as traditional opinion polling,and social media analysis.However,traditional polling approaches often struggle to capture the intricate nuances of voter sentiment at local levels,resulting in a limited depth of analysis and understanding.In light of this challenge,this study focuses on predicting elections at the state/regional level along with the country level,intending to offer a comprehensive analysis and deeper insights into the electoral process.To achieve this,the study introduces the Location-Based Election Prediction Model(LEPM),which utilizes social media data,specifically Twitter,and integrates location-aware sentiment analysis techniques at both the state/region and country levels.LEPM predicts the support and opposing strength of each political party/candidate.To determine the location of users/voters who have not disclosed their location information in tweets,the model utilizes a Voter Location Detection(VotLocaDetect)approach,which leverages recent tweets/posts.The sentiment analysis techniques employed in this study include rule-based sentiment analysis,Valence Aware Dictionary and Sentiment Reasoner(VADER)as well as transformers-based sentiment analysis such as Bidirectional Encoder Representations from Transformers(BERT),BERTweet,and Election based BERT(ElecBERT).This study uses the 2020 United States(US)Presidential Election as a case study.By applying the LEPM model to the election,the study demonstrates its ability to accurately predict outcomes in forty-one states,achieving an 0.84 accuracy rate at the state level.Moreover,at the country level,the LEPM model outperforms traditional polling results.With a low Mean Absolute Error(MAE)of 0.87,the model exhibits more precise predictions and serves as a successful alternative to conventional polls and other methodologies.Leveraging the extensive social media data,the LEPM model provides nuanced insights into voter behavior,enabling policymakers to make informed decisions and facilitating in-depth analyses of elections.The study emphasizes the importance of using social media data for reliable election prediction and offers implications for enhancing prediction accuracy and understanding voter sentiment and behavior.

Spatial Humanities and Geo-computation for Social Sciences:Advances and Applications

Humanities and Social Sciences(HSS) are undergoing the transformation of spatialization and quantification. Geo-computation, with geoinformatics(including RS: Remote Sensing;GIS: Geographical Information System;GNSS: Global Navigation Satellite System), provides effective computational and spatialization methods and tools for HSS. Spatial Humanities and Geo-computation for Social Sciences(SH&GSS) is a field coupling geo-computation, and geoinformatics, with HSS. This special issue accepted a set of contributions highlighting recent advances in methodologies and applications of SH&GSS, which are related to sentiment spatial analysis from social media data, emotional change spatial analysis from news data, spatial analysis of social media related to COVID-19, crime spatiotemporal analysis, “double evaluation” for Land Use/Land Cover(LUCC), Specially Protected Natural Areas(SPNA) analysis, editing behavior analysis of Volunteered Geographic Information(VGI), electricity consumption anomaly detection, First and Last Mile Problem(FLMP) of public transport, and spatial interaction network analysis for crude oil trade network. Based on these related researches, we aim to present an overview of SH&GSS, and propose some future research directions for SH&HSS.

Disaster Impacts Surveillance from Social Media with Topic Modeling and Feature Extraction: Case of Hurricane Harvey

Twitter can supply useful information on infrastructure impacts to the emergency managers during major disasters,but it is time consuming to filter through many irrelevant tweets.Previous studies have identified the types of messages that can be found on social media during disasters,but few solutions have been proposed to efficiently extract useful ones.We present a framework that can be applied in a timely manner to provide disaster impact information sourced from social media.The framework is tested on a well-studied and data-rich case of Hurricane Harvey.The procedures consist of filtering the raw Twitter data based on keywords,location,and tweet attributes,and then applying the latent Dirichlet allocation(LDA) to separate the tweets from the disaster affected area into categories(topics) useful to emergency managers.The LDA revealed that out of 24 topics found in the data,nine were directly related to disaster impacts-for example,outages,closures,flooded roads,and damaged infrastructure.Features such as frequent hashtags,mentions,URLs,and useful images were then extracted and analyzed.The relevant tweets,along with useful images,were correlated at the county level with flood depth,distributed disaster aid(damage),and population density.Significant correlations were found between the nine relevant topics and population density but not flood depth and damage,suggesting that more research into the suitability of social media data for disaster impacts modeling is needed.The results from this study provide baseline information for such efforts in the future.

Temporal pattern mining from user-generated content

Faster internet, IoT, and social media have reformed the conventional web into a collaborative web resulting in enormous user-generated content. Several studies are focused on such content;however, they mainly focus on textual data, thus undermining the importance of metadata. Considering this gap, we provide a temporal pattern mining framework to model and utilize user-generated content's metadata. First, we scrap 2.1 million tweets from Twitter between Nov-2020 to Sep-2021 about 100 hashtag keywords and present these tweets into 100 User-Tweet-Hashtag (UTH) dynamic graphs. Second, we extract and identify four time-series in three timespans (Day, Hour, and Minute) from UTH dynamic graphs. Lastly, we model these four time-series with three machine learning algorithms to mine temporal patterns with the accuracy of 95.89%, 93.17%, 90.97%, and 93.73%, respectively. We demonstrate that user-generated content's metadata contains valuable information, which helps to understand the users' collective behavior and can be beneficial for business and research. Dataset and codes are publicly available;the link is given in the dataset section.

Automatic Surveillance of Pandemics Using Big Data and Text Mining

COVID-19 disease is spreading exponentially due to the rapid transmission of the virus between humans.Different countries have tried different solutions to control the spread of the disease,including lockdowns of countries or cities,quarantines,isolation,sanitization,and masks.Patients with symptoms of COVID-19 are tested using medical testing kits;these tests must be conducted by healthcare professionals.However,the testing process is expensive and time-consuming.There is no surveillance system that can be used as surveillance framework to identify regions of infected individuals and determine the rate of spread so that precautions can be taken.This paper introduces a novel technique based on deep learning(DL)that can be used as a surveillance system to identify infected individuals by analyzing tweets related to COVID-19.The system is used only for surveillance purposes to identify regions where the spread of COVID-19 is high;clinical tests should then be used to test and identify infected individuals.The system proposed here uses recurrent neural networks(RNN)and word-embedding techniques to analyze tweets and determine whether a tweet provides information about COVID-19 or refers to individuals who have been infected with the virus.The results demonstrate that RNN can conduct this analysis more accurately than other machine learning(ML)algorithms.