Scaling in words on Twitter

A social media Data-Driven analysis for transport policy response to the COVID-19 pandemic outbreak in Wuhan, China

Non-pharmacological interventions (NPI) such as social distancing and lockdown are essential in preventing and controlling emerging pandemic outbreaks. Many countries worldwide implemented lockdowns during the COVID-19 outbreaks. However, due to the lack of prior experience and knowledge about the pandemic, it is challenging to deal with short-term polices decision-making due to the highly stochastic and dynamic nature of the COVID-19. Thus, there is a need for the exploration of policy decision analysis to help agencies to adjust their current policies and adopt quickly. In this study, an analytical methodology is developed to analysis urban transport policy response for pandemic control based on social media data. Compared to traditional surveys or interviews, social media can provide timely data based on the feedback from public in terms of public demands, opinions, and acceptance of policy implementations. In particular, a sentiment-aware pre-trained language model is fine-tuned for sentiment analysis of policy. The Latent Dirichlet Allocation (LDA) model is used to classify documents, e.g., posts collected from social media, into specific topics in an unsupervised manner. Then, entropy weights method (EWM) is used to extract public policy demands based on the classified topics. Meanwhile, a Jaccard distance-based approach is proposed to conduct the response analysis of policy adjustments. A retrospective analysis of transport policies during the COVID-19 pandemic in Wuhan, China is presented using the developed methodology. The results show that the developed policymaking support methodology can be an effective tool to evaluate the acceptance of anti-pandemic policies from the public's perspective, to assess the balance between policies and people's demands, and to further perform the response analysis of a series of policy adjustments based on online feedback.

Storywrangler: A massive exploratorium for sociolinguistic, cultural, socioeconomic, and political timelines using Twitter

In real time, Twitter strongly imprints world events, popular culture, and the day-to-day, recording an ever-growing compendium of language change. Vitally, and absent from many standard corpora such as books and news archives, Twitter also encodes popularity and spreading through retweets. Here, we describe Storywrangler, an ongoing curation of over 100 billion tweets containing 1 trillion 1-grams from 2008 to 2021. For each day, we break tweets into 1-, 2-, and 3-grams across 100+ languages, generating frequencies for words, hashtags, handles, numerals, symbols, and emojis. We make the dataset available through an interactive time series viewer and as downloadable time series and daily distributions. Although Storywrangler leverages Twitter data, our method of tracking dynamic changes in n-grams can be extended to any temporally evolving corpus. Illustrating the instrument's potential, we present example use cases including social amplification, the sociotechnical dynamics of famous individuals, box office success, and social unrest.

Uncovering the behaviour of road accidents in urban areas

Different patterns in the incidence of road accidents are revealed when considering areas with increased levels of urbanization. To understand these patterns, road accident data from England and Wales is explored. In particular, the data are used to (i) generate time series for comparison of the incidence of road accidents in urban as opposed to rural areas, (ii) analyse the relationship between the number of road accidents and the population size of a set of urban areas, and (iii) model the likelihood of suffering an accident in an urban area and its dependence with population size. It is observed that minor and serious accidents are more frequent in urban areas, whereas fatal accidents are more likely in rural areas. It is also shown that, generally, the number of accidents in an urban area depends on population size superlinearly, with this superlinear behaviour becoming stronger for lower degrees of severity. Finally, given an accident in an urban area, the probability that the accident is fatal or serious decreases with population size and the probability that it is minor, increases sublinearly. These findings promote the question as to why such behaviours exist, the answer to which will lead to more sustainable urban policies.