A Statistical Approach for Studying the Spatio-Temporal Distribution of Geolocated Tweets in Urban Environments

Abstract

An in-depth descriptive approach to the dynamics of the urban population is fundamental as a first step towards promoting effective planning and designing processes in cities. Understanding the behavioral aspects of human activities can contribute to their effective management and control. We present a framework, based on statistical methods, for studying the spatio-temporal distribution of geolocated tweets as a proxy for where and when people carry out their activities. We have evaluated our proposal by analyzing the distribution of collected geolocated tweets over a two-week period in the summer of 2017 in Lisbon, London, and Manhattan. Our proposal considers a negative binomial regression analysis for the time series of counts of tweets as a first step. We further estimate a functional principal component analysis of second-order summary statistics of the hourly spatial point patterns formed by the locations of the tweets. Finally, we find groups of hours with a similar spatial arrangement of places where humans develop their activities through hierarchical clustering over the principal scores. Social media events are found to show strong temporal trends such as seasonal variation due to the hour of the day and the day of the week in addition to autoregressive schemas. We have also identified spatio-temporal patterns of clustering, i.e., groups of hours of the day that present a similar spatial distribution of human activities.

A joint Bayesian spatiotemporal risk prediction model of COVID-19 incidence, IC admission, and death with application to Sweden

“…The R-INLA package also provides goodness-of-fit statistics, notably the deviance information criterion (DIC) (Spiegelhalter et al 2002 ) and the Watanabe Akaike information criterion (WAIC) (Watanabe 2010 ). It also yields the marginal predictive likelihood (MPL) (Dey et al 1997 ), mean absolute error (MAE), root mean square error (RMSE) (Pal 2017 ), and Pearson correlation coefficient ( r ) (Santa et al 2019 ), which are appropriate statistics for prediction performance evaluation.…”

Emotive facets of place meet urban analytics

“…The addition of time in place‐based methods allows the study of the evolution of emotions framed in and from places. In urban analytics, longitudinal studies, based on proxy data (e.g., Twitter, smartcard data), of individuals' movements across the city have shown how to deal with mobility models and networks (Barbosa et al, 2018), space–time prisms (Senaratne et al, 2017), and detect mobility patterns over a prolonged time period (Kandt & Batty, 2021; Kandt & Leak, 2019; Santa et al, 2019). Yet, the use of objective data has taken a rise with the use of GPS sensors in smartphones and wearables; for example, speed, travel time and delay for intersections and road segments (Strauss & Miranda‐Moreno, 2017), traffic and road condition estimation (Bhoraskar et al, 2012), pothole detection (Xue et al, 2016), and fall detection in urban contexts (Lee et al, 2018).…”

Bayesian spatiotemporal forecasting and mapping of COVID‐19 risk with application to West Java Province, Indonesia

“…Furthermore, the INLA package yields the deviance information criterion (DIC) (Spiegelhalter et al, 2002) and Watanabe–Akaike information criterion (WAIC) (Watanabe, 2010) as goodness‐of‐fit statistics. INLA also yields the marginal predictive‐likelihood (MPL) (Dey et al, 1997), mean absolute error (MAE), root mean squared error (RMSE) (Pal, 2017), and Pearson correlation coefficient ( r ) (Santa et al, 2019), which are appropriate statistics for prediction performance evaluation as such (if all the observations are used to estimate the model), and for evaluation based on cross‐validation.…”