arXiv subjects

Satish V. Ukkusuri

Publications and source records attributed to Satish V. Ukkusuri.

48 records · Page 3Linked to original sources

arXiv2019

Predicting Evacuation Decisions using Representations of Individuals' Pre-Disaster Web Search Behavior

Predicting the evacuation decisions of individuals before the disaster strikes is crucial for planning first response strategies. In addition to the studies on post-disaster analysis of evacuation behavior, there are various works that attempt to predict the evacuation decisions beforehand. Most of these predictive methods, however, require real time location data for calibration, which are becoming much harder to obtain due to the rising privacy concerns. Meanwhile, web search queries of anonymous users have been collected by web companies. Although such data raise less privacy concerns, they have been under-utilized for various applications. In this study, we investigate whether web search data observed prior to the disaster can be used to predict the evacuation decisions. More specifically, we utilize a "session-based query encoder" that learns the representations of each user's web search behavior prior to evacuation. Our proposed approach is empirically tested using web search data collected from users affected by a major flood in Japan. Results are validated using location data collected from mobile phones of the same set of users as ground truth. We show that evacuation decisions can be accurately predicted (84%) using only the users' pre-disaster web search data as input. This study proposes an alternative method for evacuation prediction that does not require highly sensitive location data, which can assist local governments to prepare effective first response strategies.

cs.SI↗

arXiv2019

Universality of population recovery patterns after disasters

Despite the rising importance of enhancing community resilience to disasters, our understanding on how communities recover from catastrophic events is limited. Here we study the population recovery dynamics of disaster affected regions by observing the movements of over 2.5 million mobile phone users across three countries before, during and after five major disasters. We find that, although the regions affected by the five disasters have significant differences in socio-economic characteristics, we observe a universal recovery pattern where displaced populations return in an exponential manner after all disasters. Moreover, the heterogeneity in initial and long-term displacement rates across communities across the three countries were explained by a set of key universal factors including the community's median income level, population size, housing damage rate, and the connectedness to other cities. These universal properties of recovery dynamics extracted from large scale evidence could impact efforts on urban resilience and sustainability across various disciplines.

physics.soc-ph↗

arXiv2019

Influencing factors that determine the usage of the crowd-shipping services

The objective of this study is to understand how senders choose shipping services for different products, given the availability of both emerging crowd-shipping (CS) and traditional carriers in a logistics market. Using data collected from a US survey, Random Utility Maximization (RUM) and Random Regret Minimization (RRM) models have been employed to reveal factors that influence the diversity of decisions made by senders. Shipping costs, along with additional real-time services such as courier reputations, tracking info, e-notifications, and customized delivery time and location, have been found to have remarkable impacts on senders' choices. Interestingly, potential senders were willing to pay more to ship grocery items such as food, beverages, and medicines by CS services. Moreover, the real-time services have low elasticities, meaning that only a slight change in those services will lead to a change in sender-behavior. Finally, data-science techniques were used to assess the performance of the RUM and RRM models and found to have similar accuracies. The findings from this research will help logistics firms address potential market segments, prepare service configurations to fulfill senders' expectations, and develop effective business operations strategies.

econ.GN↗

arXiv2018

Modeling Urban Taxi Services with e-hailings: A Queueing Network Approach

The rise of e-hailing taxis has significantly altered urban transportation and resulted in a competitive taxi market with both traditional street-hailing and e-hailing taxis. The new mobility services provide similar door-to-door rides as the traditional one and there is competition across these various services. In this study, we propose an innovative queueing network model for the competitive taxi market and capture the interactions not only within the taxi market but also between the taxi market and urban road system. An example is designed based on data from New York City. Numerical results show that the proposed modeling structure, together with the corresponding stationary limits, can capture dynamics within high demand areas using multiple data sources. Overall, this study shows how the queueing network approach can measure both the taxi and urban road system performance at an aggregate level. The model can be used to estimate not only the waiting/searching time during passenger-vehicle matching but also the delays in the urban road network. Furthermore, the model can be generalized to study the control and management of taxi markets.

stat.AP↗

arXiv2018

Joint modeling of evacuation departure and travel times in hurricanes

Hurricanes are costly natural disasters periodically faced by households in coastal and to some extent, inland areas. A detailed understanding of evacuation behavior is fundamental to the development of efficient emergency plans. Once a household decides to evacuate, a key behavioral issue is the time at which individuals depart to reach their destination. An accurate estimation of evacuation departure time is useful to predict evacuation demand over time and develop effective evacuation strategies. In addition, the time it takes for evacuees to reach their preferred destinations is important. A holistic understanding of the factors that affect travel time is useful to emergency officials in controlling road traffic and helps in preventing adverse conditions like traffic jams. Past studies suggest that departure time and travel time can be related. Hence, an important question arises whether there is an interdependence between evacuation departure time and travel time? Does departing close to the landfall increases the possibility of traveling short distances? Are people more likely to depart early when destined to longer distances? In this study, we present a model to jointly estimate departure and travel times during hurricane evacuations. Empirical results underscore the importance of accommodating an inter-relationship among these dimensions of evacuation behavior. This paper also attempts to empirically investigate the influence of social ties of individuals on joint estimation of evacuation departure and travel times. Survey data from Hurricane Sandy is used for computing empirical results. Results indicate significant role of social networks in addition to other key factors on evacuation departure and travel times during hurricanes.

stat.AP↗

arXiv2018

User equilibrium with a policy-based link transmission model for stochastic time-dependent traffic networks

Non-recurrent congestion is a major problem in traffic networks that causes unexpected delays during travels. In such a scenario, it is preferable to use adaptive paths or policies where next link decisions on reaching junctions are continuously adapted based on the information gained with time. In this paper, we study a traffic assignment problem in stochastic time-dependent networks. The problem is modeled as a fixed-point problem and existence of the equilibrium solution is discussed. We iteratively solve the problem using the method of successive averages (MSA). A novel network loading model inspired from Link transmission model is developed that accepts policies as inputs for solving the problem. This network loading model is different from the existing network loading models that take predefined paths for input flows. We demonstrate through numerical tests that solving traffic assignment problem with the proposed loading modeling scheme is more efficient as compared to solving the problem using path-based network loading models.

eess.SY↗

arXiv2018

Selectivity correction in discrete-continuous models for the willingness to work as crowd-shippers and travel time tolerance

The objective of this study is to understand the different behavioral considerations that govern the choice of people to engage in a crowd-shipping market. Using novel data collected by the researchers in the US, we develop discrete-continuous models. A binary logit model has been used to estimate crowd-shippers' willingness to work, and an ordinary least-square regression model has been employed to calculate crowd-shippers' maximum tolerance for shipping and delivery times. A selectivity-bias term has been included in the model to correct for the conditional relationships of the crowd-shipper's willingness to work and their maximum travel time tolerance. The results show socio-demographic characteristics (e.g. age, gender, race, income, and education level), transporting freight experience, and number of social media usages significant influence the decision to participate in the crowd-shipping market. In addition, crowd-shippers pay expectations were found to be reasonable and concurrent with the literature on value-of-time. Findings from this research are helpful for crowd-shipping companies to identify and attract potential shippers. In addition, an understanding of crowd-shippers - their behaviors, perceptions, demographics, pay expectations, and in which contexts they are willing to divert from their route - are valuable to the development of business strategies such as matching criteria and compensation schemes for driver-partners.

econ.GN↗

arXiv2018

Crowd-shipping services for last mile delivery: analysis from survey data in two countries

The e-commerce boom has led to overwhelming demand for personalized delivery services. Accordingly, various start-ups and tech companies provide crowd-shipping services that aim to be more efficient and effective than traditional logistics options. These services are fueled by technological innovation, improved internet infrastructure, and increased smartphone use. However, the field of on-demand delivery faces several challenges, including specified pickup and delivery times and locations. Therefore, market demand and prospective crowd-shipper supply must be well understood to ensure industry success. This research analyzes current and future shipping behaviors, as well as potential employees' willingness to work (WTW) as crowd-shippers. Revealed and stated preference survey questionnaires were designed. The surveys were implemented in Vietnam and the US. This descriptive study makes use of the survey data sets to understand the behavior of requesters and potential crowd-shippers in the logistics market and assumes that crowd-sourced delivery is available. The results show requesters' various behaviors and expectations as well as prospective crowd-shippers' WTW in the two countries. The results can be used to recruit potential crowd-shippers and create business strategies that match requesters' and potential crowd-shippers' expectations.

physics.soc-ph↗

arXiv2017

Crisis Communication Patterns in Social Media during Hurricane Sandy

Hurricane Sandy was one of the deadliest and costliest of hurricanes over the past few decades. Many states experienced significant power outage, however many people used social media to communicate while having limited or no access to traditional information sources. In this study, we explored the evolution of various communication patterns using machine learning techniques and determined user concerns that emerged over the course of Hurricane Sandy. The original data included ~52M tweets coming from ~13M users between October 14, 2012 and November 12, 2012. We run topic model on ~763K tweets from top 4,029 most frequent users who tweeted about Sandy at least 100 times. We identified 250 well-defined communication patterns based on perplexity. Conversations of most frequent and relevant users indicate the evolution of numerous storm-phase (warning, response, and recovery) specific topics. People were also concerned about storm location and time, media coverage, and activities of political leaders and celebrities. We also present each relevant keyword that contributed to one particular pattern of user concerns. Such keywords would be particularly meaningful in targeted information spreading and effective crisis communication in similar major disasters. Each of these words can also be helpful for efficient hash-tagging to reach target audience as needed via social media. The pattern recognition approach of this study can be used in identifying real time user needs in future crises.

cs.SI↗

arXiv2017

Understanding Information Spreading in Social Media during Hurricane Sandy: User Activity and Network Properties

Many people use social media to seek information during disasters while lacking access to traditional information sources. In this study, we analyze Twitter data to understand information spreading activities of social media users during hurricane Sandy. We create multiple subgraphs of Twitter users based on activity levels and analyze network properties of the subgraphs. We observe that user information sharing activity follows a power-law distribution suggesting the existence of few highly active nodes in disseminating information and many other nodes being less active. We also observe close enough connected components and isolates at all levels of activity, and networks become less transitive, but more assortative for larger subgraphs. We also analyze the association between user activities and characteristics that may influence user behavior to spread information during a crisis. Users become more active in spreading information if they are centrally placed in the network, less eccentric, and have higher degrees. Our analysis provides insights on how to exploit user characteristics and network properties to spread information or limit the spreading of misinformation during a crisis event.

cs.SI↗

arXiv2017

Analyzing Social Interaction Networks from Twitter for Planned Special Events

The complex topology of real networks allows its actors to change their functional behavior. Network models provide better understanding of the evolutionary mechanisms being accountable for the growth of such networks by capturing the dynamics in the ways network agents interact and change their behavior. Considerable amount of research efforts is required for developing novel network modeling techniques to understand the structural properties such networks, reproducing similar properties based on empirical evidence, and designing such networks efficiently. First, we demonstrate how to construct social interaction networks using social media data and then present the key findings obtained from the network analytics. We analyze the characteristics and growth of such interaction networks, examine the network properties and derive important insights based on the theories of network science literature. We also discuss the application of such networks as a useful tool to effectively disseminate targeted information during planned special events. We observed that the degree-distributions of such networks follow power-law that is indicative of the existence of fewer nodes in the network with higher levels of interactions, and many other nodes with less interactions. While the network elements and average user degree grow linearly each day, densities of such networks tend to become zero. Largest connected components exhibit higher connectivity (density) when compared with the whole graph. Network radius and diameter become stable over time evidencing the small-world property. We also observe increased transitivity and higher stability of the power-law exponents as the networks grow. Data is specific to the Purdue University community and two large events, namely Purdue Day of Giving and Senator Bernie Sanders' visit to Purdue University as part of Indiana Primary Election 2016.

cs.SI↗

arXiv2017

Joint Inference of User Community and Interest Patterns in Social Interaction Networks

Online social media have become an integral part of our social beings. Analyzing conversations in social media platforms can lead to complex probabilistic models to understand social interaction networks. In this paper, we present a modeling approach for characterizing social interaction networks by jointly inferring user communities and interests based on social media interactions. We present several pattern inference models: i) Interest pattern model (IPM) captures population level interaction topics, ii) User interest pattern model (UIPM) captures user specific interaction topics, and iii) Community interest pattern model (CIPM) captures both community structures and user interests. We test our methods on Twitter data collected from Purdue University community. From our model results, we observe the interaction topics and communities related to two big events within Purdue University community, namely Purdue Day of Giving and Senator Bernie Sanders' visit to Purdue University as part of Indiana Primary Election 2016. Constructing social interaction networks based on user interactions accounts for the similarity of users' interactions on various topics of interest and indicates their community belonging further beyond connectivity. We observed that the degree-distributions of such networks follow power-law that is indicative of the existence of fewer nodes in the network with higher levels of interactions, and many other nodes with less interactions. We also discuss the application of such networks as a useful tool to effectively disseminate specific information to the target audience towards planning any large-scale events and demonstrate how to single out specific nodes in a given community by running network algorithms.

cs.SI↗