Search

Introduction: Wearable devices are rapidly improving our ability to observe health-related processes for extended durations in an unintrusive manner. In this study, we use wearable devices to understand how the shape of the heart rate curve during sleep relates to mental health. Methods: As part of the Lived Experiences Measured Using Rings Study (LEMURS), we collected heart rate measurements using the Oura ring (Gen3) for over 25,000 sleep periods and self-reported mental health indicators from roughly 600 first-year university students in the USA during the fall semester of 2022. Using clustering techniques, we find that the sleeping heart rate curves can be broadly separated into two categories that are mainly differentiated by how far along the sleep period the lowest heart rate is reached. Results: Sleep periods characterized by reaching the lowest heart rate later during sleep are also associated with shorter deep and REM sleep and longer light sleep, but not a difference in total sleep duration. Aggregating sleep periods at the individual level, we find that consistently reaching the lowest heart rate later during sleep is a significant predictor of (1) self-reported impairment due to anxiety or depression, (2) a prior mental health diagnosis, and (3) firsthand experience in traumatic events. This association is more pronounced among females. Conclusion: Our results show that the shape of the sleeping heart rate curve, which is only weakly correlated with descriptive statistics such as the average or the minimum heart rate, is a viable but mostly overlooked metric that can help quantify the relationship between sleep and mental health.

Objective: Panic attacks are an impairing mental health problem that affects 11% of adults every year. Current criteria describe them as occurring without warning, despite evidence suggesting individuals can often identify attack triggers. We aimed to prospectively explore qualitative and quantitative factors associated with the onset of panic attacks. Results: Of 87 participants, 95% retrospectively identified a trigger for their panic attacks. Worse individually reported mood and state-level mood, as indicated by Twitter ratings, were related to greater likelihood of next-day panic attack. In a subsample of participants who uploaded their wearable sensor data (n = 32), louder ambient noise and higher resting heart rate were related to greater likelihood of next-day panic attack. Conclusions: These promising results suggest that individuals who experience panic attacks may be able to anticipate their next attack which could be used to inform future prevention and intervention efforts.

Abstract Complex systems often comprise many kinds of components which vary over many orders of magnitude in size: Populations of cities in countries, individual and corporate wealth in economies, species abundance in ecologies, word frequency in natural language, and node degree in complex networks. Here, we introduce ‘allotaxonometry’ along with ‘rank-turbulence divergence’ (RTD), a tunable instrument for comparing any two ranked lists of components. We analytically develop our rank-based divergence in a series of steps, and then establish a rank-based allotaxonograph which pairs a map-like histogram for rank-rank pairs with an ordered list of components according to divergence contribution. We explore the performance of rank-turbulence divergence, which we view as an instrument of ‘type calculus’, for a series of distinct settings including: Language use on Twitter and in books, species abundance, baby name popularity, market capitalization, performance in sports, mortality causes, and job titles. We provide a series of supplementary flipbooks which demonstrate the tunability and storytelling power of rank-based allotaxonometry.

Abstract While quantitative methods have been used to examine changes in word usage in books, studies have focused on overall trends, such as the shapes of narratives, which are independent of book length. We instead look at how words change over the course of a book as a function of the number of words, rather than the fraction of the book, completed at any given point; we define this measure as “cumulative word-time”. Using ousiometrics, a reinterpretation of the valence–arousal–dominance framework of meaning obtained from semantic differentials, we convert text into time series of power and danger scores, with time corresponding to cumulative word-time. Each time series is then decomposed using empirical mode decomposition into a sum of constituent oscillatory modes and a non-oscillatory trend. By comparing the decomposition of the original power and danger time series with those derived from shuffled text, we find that shorter books exhibit only a general trend, while longer books have fluctuations in addition to the general trend. These fluctuations typically have a period of a few thousand words regardless of the book length or library classification code but vary depending on the content and structure of the book. Our findings suggest that, in the ousiometric sense, longer books are not expanded versions of shorter books, but rather are more similar in structure to a concatenation of shorter texts. Further, they are consistent with editorial practices that require longer texts to be broken down into sections, such as chapters. Our method also provides a data-driven denoising approach that works for texts of various lengths, in contrast to the more traditional approach of using large window sizes that may inadvertently smooth out relevant information, especially for shorter texts. Altogether, these results open up avenues for future work in computational literary analysis, particularly the possibility of measuring a basic unit of narrative.

Abstract Online misogyny has become a fixture in female politicians’ lives. Backlash theory suggests that it may represent a threat response prompted by female politicians’ counterstereotypical, power-seeking behaviors. We investigated this hypothesis by analyzing Twitter references to Hillary Clinton before, during, and after her presidential campaign. We collected a corpus of over 9 million tweets from 2014 to 2018 that referred to Hillary Clinton, and employed an interrupted time series analysis on the relative frequency of misogynistic language within the corpus. Prior to 2015, the level of misogyny associated with Clinton decreased over time, but this trend reversed when she announced her presidential campaign. During the campaign, misogyny steadily increased and only plateaued after the election, when the threat of her electoral success had subsided. These findings are consistent with the notion that online misogyny towards female political nominees is a form of backlash prompted by their ambition for power in the political arena.

Abstract We explore the relationship between context and happiness scores in political tweets using word co-occurrence networks, where nodes in the network are the words, and the weight of an edge is the number of tweets in the corpus for which the two connected words co-occur. In particular, we consider tweets with hashtags #imwithher and #crookedhillary, both relating to Hillary Clinton’s presidential bid in 2016. We then analyze the network properties in conjunction with the word scores by comparing with null models to separate the effects of the network structure and the score distribution. Neutral words are found to be dominant and most words, regardless of polarity, tend to co-occur with neutral words. We do not observe any score homophily among positive and negative words. However, when we perform network backboning, community detection results in word groupings with meaningful narratives, and the happiness scores of the words in each group correspond to its respective theme. Thus, although we observe no clear relationship between happiness scores and co-occurrence at the node or edge level, a community-centric approach can isolate themes of competing sentiments in a corpus.

The murder of George Floyd by police in May 2020 sparked international protests and brought unparalleled levels of attention to the Black Lives Matter movement. As we show, his death set record levels of activity and amplification on Twitter, prompted the saddest day in the platform’s history, and caused his name to appear among the ten most frequently used phrases in a day, where he is the only individual to have ever received that level of attention who was not known to the public earlier that same week. Importantly, we find that the Black Lives Matter movement’s rhetorical strategy to connect and repeat the names of past Black victims of police violence—foregrounding racial injustice as an ongoing pattern rather than a singular event—was exceptionally effective following George Floyd’s death: attention given to him extended to over 185 prior Black victims, more than other past moments in the movement’s history. We contextualize this rising tide of attention among 12 years of racial justice activism on Twitter, demonstrating how activists and allies have used attention and amplification as a recurring tactic to lift and memorialize the names of Black victims of police violence. Our results show how the Black Lives Matter movement uses social media to center past instances of police violence at an unprecedented scale and speed, while still advancing the racial justice movement’s longstanding goal to “say their names.”

Abstract Sleep loss has been linked to heart disease, diabetes, cancer, and an increase in accidents, all of which are among the leading causes of death in the United States. Population-scale sleep studies have the potential to advance public health by helping to identify at-risk populations, changes in collective sleep patterns, and to inform policy change. Prior research suggests other kinds of health indicators such as depression and obesity can be estimated using social media activity. However, the inability to effectively measure collective sleep with publicly available data has limited large-scale academic studies. Here, we investigate the passive estimation of sleep loss through a proxy analysis of Twitter activity profiles. We use “Spring Forward” events, which occur at the beginning of Daylight Savings Time in the United States, as a natural experimental condition to estimate spatial differences in sleep loss across the United States. On average, peak Twitter activity occurs 15 to 30 min later on the Sunday following Spring Forward. By Monday morning however, activity curves are realigned with the week before, suggesting that the window of sleep opportunity is compressed in Twitter data, revealing Spring Forward behavioral change.

Abstract Working from a dataset of 118 billion messages running from the start of 2009 to the end of 2019, we identify and explore the relative daily use of over 150 languages on Twitter. We find that eight languages comprise 80% of all tweets, with English, Japanese, Spanish, Arabic, and Portuguese being the most dominant. To quantify social spreading in each language over time, we compute the ‘contagion ratio’: The balance of retweets to organic messages. We find that for the most common languages on Twitter there is a growing tendency, though not universal, to retweet rather than share new content. By the end of 2019, the contagion ratios for half of the top 30 languages, including English and Spanish, had reached above 1—the naive contagion threshold. In 2019, the top 5 languages with the highest average daily ratios were, in order, Thai (7.3), Hindi, Tamil, Urdu, and Catalan, while the bottom 5 were Russian, Swedish, Esperanto, Cebuano, and Finnish (0.26). Further, we show that over time, the contagion ratios for most common languages are growing more strongly than those of rare languages.

Abstract A common task in computational text analyses is to quantify how two corpora differ according to a measurement like word frequency, sentiment, or information content. However, collapsing the texts’ rich stories into a single number is often conceptually perilous, and it is difficult to confidently interpret interesting or unexpected textual patterns without looming concerns about data artifacts or measurement validity. To better capture fine-grained differences between texts, we introduce generalized word shift graphs, visualizations which yield a meaningful and interpretable summary of how individual words contribute to the variation between two texts for any measure that can be formulated as a weighted average. We show that this framework naturally encompasses many of the most commonly used approaches for comparing texts, including relative frequencies, dictionary scores, and entropy-based measures like the Kullback–Leibler and Jensen–Shannon divergences. Through a diverse set of case studies ranging from presidential speeches to tweets posted in urban green spaces, we demonstrate how generalized word shift graphs can be flexibly applied across domains for diagnostic investigation, hypothesis generation, and substantive interpretation. By providing a detailed lens into textual shifts between corpora, generalized word shift graphs help computational social scientists, digital humanists, and other text analysis practitioners fashion more robust scientific narratives.

Sentiment-aware intelligent systems are essential to a wide array of applications. These systems are driven by language models which broadly fall into two paradigms: Lexicon-based and contextual. Although recent contextual models are increasingly dominant, we still see demand for lexicon-based models because of their interpretability and ease of use. For example, lexicon-based models allow researchers to readily determine which words and phrases contribute most to a change in measured sentiment. A challenge for any lexicon-based approach is that the lexicon needs to be routinely expanded with new words and expressions. Here, we propose two models for automatic lexicon expansion. Our first model establishes a baseline employing a simple and shallow neural network initialized with pre-trained word embeddings using a non-contextual approach. Our second model improves upon our baseline, featuring a deep Transformer-based network that brings to bear word definitions to estimate their lexical polarity. Our evaluation shows that both models are able to score new words with a similar accuracy to reviewers from Amazon Mechanical Turk, but at a fraction of the cost.

The COVID-19 pandemic disrupted the mobility patterns of a majority of Americans beginning in March 2020. Despite the beneficial, socially distanced activity offered by outdoor recreation, confusing and contradictory public health messaging complicated access to natural spaces. Working with a dataset comprising the locations of roughly 50 million distinct mobile devices in 2019 and 2020, we analyze weekly visitation patterns for 8,135 parks across the United States. Using Bayesian inference, we identify regions that experienced a substantial change in visitation in the first few weeks of the pandemic. We find that regions that did not exhibit a change were likely to have smaller populations, and to have voted more republican than democrat in the 2020 elections. Our study contributes to a growing body of literature using passive observations to explore who benefits from access to nature.

The U.S. stock market is one of the largest and most complex marketplaces in the global financial system. Over the past several decades, this market has evolved at multiple structural and temporal scales. New exchanges became active, and others stopped trading, regulations have been introduced and adapted, and technological innovations have pushed the pace of trading activity to blistering speeds. These developments have supported the growth of a rich machine-trading ecology that leads to qualitative differences in trading behavior at human and machine time scales. We conduct a longitudinal analysis of comprehensive market data to quantify nonstationary dynamics throughout this system. We quantify the relationship between fluctuations in the number of active trading venues and realized opportunity costs experienced by market participants. We find that information asymmetries, in the form of quote dislocations, predict market-wide volatility indicators. Lastly, we uncover multiple micro-to-macro level pathways, including those exhibiting evidence of self-organized criticality.

Abstract We introduce a qualitative, shape-based, timescale-independent time-domain transform used to extract local dynamics from sociotechnical time series—termed the Discrete Shocklet Transform (DST)—and an associated similarity search routine, the Shocklet Transform And Ranking (STAR) algorithm, that indicates time windows during which panels of time series display qualitatively-similar anomalous behavior. After distinguishing our algorithms from other methods used in anomaly detection and time series similarity search, such as the matrix profile, seasonal-hybrid ESD, and discrete wavelet transform-based procedures, we demonstrate the DST’s ability to identify mechanism-driven dynamics at a wide range of timescales and its relative insensitivity to functional parameterization. As an application, we analyze a sociotechnical data source (usage frequencies for a subset of words on Twitter) and highlight our algorithms’ utility by using them to extract both a typology of mechanistic local dynamics and a data-driven narrative of socially-important events as perceived by English-language Twitter.

Abstract With more people living in cities, we are witnessing a decline in exposure to nature. A growing body of research has demonstrated an association between nature contact and improved mood. Here, we used Twitter and the Hedonometer, a world analysis tool, to investigate how sentiment, or the estimated happiness of the words people write, varied before, during, and after visits to San Francisco's urban park system. We found that sentiment was substantially higher during park visits and remained elevated for several hours following the visit. Leveraging differences in vegetative cover across park types, we explored how different types of outdoor public spaces may contribute to subjective well‐being. Tweets during visits to Regional Parks, which are greener and have greater vegetative cover, exhibited larger increases in sentiment than tweets during visits to Civic Plazas and Squares. Finally, we analysed word frequencies to explore several mechanisms theorized to link nature exposure with mental and cognitive benefits. Negation words such as ‘no’, ‘not’, and ‘don't’ decreased in frequency during visits to urban parks. These results can be used by urban planners and public health officials to better target nature contact recommendations for growing urban populations. A free Plain Language Summary can be found within the Supporting Information of this article.

Measuring the specific kind, temporal ordering, diversity, and turnover rate of stories surrounding any given subject is essential to developing a complete reckoning of that subject’s historical impact. Here, we use Twitter as a distributed news and opinion aggregation source to identify and track the dynamics of the dominant day-scale stories around Donald Trump, the 45th President of the United States. Working with a data set comprising around 20 billion 1-grams, we first compare each day’s 1-gram and 2-gram usage frequencies to those of a year before, to create day- and week-scale timelines for Trump stories for 2016–2021. We measure Trump’s narrative control, the extent to which stories have been about Trump or put forward by Trump. We then quantify story turbulence and collective chronopathy—the rate at which a population’s stories for a subject seem to change over time. We show that 2017 was the most turbulent overall year for Trump. In 2020, story generation slowed dramatically during the first two major waves of the COVID-19 pandemic, with rapid turnover returning first with the Black Lives Matter protests following George Floyd’s murder and then later by events leading up to and following the 2020 US presidential election, including the storming of the US Capitol six days into 2021. Trump story turnover for 2 months during the COVID-19 pandemic was on par with that of 3 days in September 2017. Our methods may be applied to any well-discussed phenomenon, and have potential to enable the computational aspects of journalism, history, and biography.

As the most visible face of health expertise to the general public, health agencies have played a central role in alerting the public to the emerging COVID-19 threat, providing guidance for protective action, motivating compliance with health directives, and combating misinformation. Social media platforms such as Twitter have been a critical tool in this process, providing a communication channel that allows both rapid dissemination of messages to the public at large and individual-level engagement. Message dissemination and amplification is a necessary precursor to reaching audiences, both online and off, as well as inspiring action. Therefore, it is valuable for organizational risk communication to identify strategies and practices that may lead to increased message passing among online users. In this research, we examine message features shown in prior disasters to increase or decrease message retransmission under imminent threat conditions to develop models of official risk communicators’ messages shared online from February 1, 2020-April 30, 2020. We develop a lexicon of keywords associated with risk communication about the pandemic response, then use automated coding to identify message content and message structural features. We conduct chi-square analyses and negative binomial regression modeling to identify the strategies used by official risk communicators that respectively increase and decrease message retransmission. Findings show systematic changes in message strategies over time and identify key features that affect message passing, both positively and negatively. These results have the potential to aid in message design strategies as the pandemic continues, or in similar future events.

A simplified model of natural convection, similar to the Lorenz system, is compared to computational fluid dynamics simulations of a thermosyphon in order to test data assimilation (DA) methods and better understand the dynamics of convection. The thermosyphon is represented by a long time flow simulation, which serves as a reference ‘truth’. Forecasts are then made using the Lorenz-like model and synchronised to noisy and limited observations of the truth using DA. The resulting analysis is observed to infer dynamics absent from the model when using short assimilation windows. Furthermore, chaotic flow reversal occurrence and residency times in each rotational state are forecast using analysis data. Flow reversals have been successfully forecast in the related Lorenz system, as part of a perfect model experiment, but never in the presence of significant model error or unobserved variables. Finally, we provide new details concerning the fluid dynamical processes present in the thermosyphon during these flow reversals.