Geospatial Big Data Analytics
Permanent URI for this collection
Browse
Recent Submissions
Item Examining Common Factors Influencing Suburban Transit Ridership – A Case Study of Metro Stations outside Center-Shanghai(2025-01-07) Jin, ChenweiResearch have suggested that TOD (transit-oriented development) in sub-centers with transit stations is an approach to increase foot traffic and the overall vitality of the area. However, current development of sub-centers mostly follows the established TOD model in CAZs (central activity zone), and most TOD studies do not deliberately distinguish or emphasize between CAZs and sub-centers. Therefore, this study aims to examine the influence of factors in the built-up area of Shanghai's urban sub-center on TOD-ness around the station. The four main factors are accessibility to private cars, land use, inner connectivity, and functionality. A regression model was introduced to validate the factors that significantly influence transit ridership. Results show that most of factors are relative to metro usage, with a lower correlation index, indicating the extent of the impact they bring varies remarkably with the urban context. while some factors have conversed influence as in urban centers.Item Enhanced Business Location Intelligence by Forecasting Innovation Adopter Types and Predictive Space-Time Locations: An Integrated Approach(2025-01-07) Franklin, Christopher; Sreedharan, Jayashree; Bezak, VictorAn integrated temporal forecast with Bayesian predicted spatial presence of Rogers innovation diffusion adopters is introduced. A “Bass-Bayes-Spatial-Extension” (BBSE) model forms the foundation of the ensemble; with Rogers’ Innovation Adopter Behavioral Types added. We discover the simpatico relationship between Bass “Innovation Adopters”, and Rogers “Innovation Adopters”. When they are synchronized, we provide new information on WHEN, WHO and WHERE Adopter types enter the market. Demand chain management’s strategic marketing-mix decisions benefit dynamically from this parsimonious combined geospatial big data – enhanced Location Intelligence approach. The (a) Rogers’ psycho-social profiles and (b) evidence of ongoing geocoded random empirical adoptions (from actual ongoing geocoded adopter sales), iteratively update and inform their hierarchical posteriors and prior distributions. Bayesian probabilistic predictions of spatial distribution over the Census areal units-of-analysis emerge with advanced demographic and social analysis. A simple hypothetical example shows the practitioner the basic operationalization of the model. A summary and conclusion complete the paperItem Beyond the Scrimmage Line: A Geospatial Big Data Perspective on NFL Special Teams(2025-01-07) Koohikamali, Mehrdad; Dockins, ErinnThis paper uses the LA Rams as an example to explore often overlooked areas of NFL team strategy, focusing on the draft and special teams. The NFL draft plays a vital role in securing a competitive edge, while special teams, although less highlighted, play a crucial role in scoring. However, these areas are rarely analyzed in depth. By leveraging extensive data from the NFL's 'Next Gen Stats,' this study examines both draft performance and special team plays. We developed three models—a classification, a regression, and a hotspot analysis—to better understand player performance and identify strategic improvements. The results aim to improve decision-making and strategy for NFL teams, offering new insights into player roles, the effects of player positioning, and key areas on the field where successful plays occur.Item Improving Loan Servicing in Microfinance: A Mobility-Network Approach(2025-01-07) Tan, Tianhui; Phan, TuanMicrofinance organizations have emerged as a lasting solution to the financial exclusion problem in several emerging economies. With a rapidly increasing penetration of mobile phones, and Internet services in these countries, organizations are now exploring novel sources of information for servicing micro-loans and assessing creditworthiness. In this paper, we emphasize the importance of leveraging the network of borrowers in predicting defaults for loan servicing and collection. Specifically, we propose a modeling technique that uses the mobility pattern of borrowers, following the loan approval, to create representation of co-located network among the borrowers. Through collaboration with a large consumer finance marketplace in China, we illustrate the statistical power of such network representations at predicting loan repayment behavior, for a selected sample of borrowers. Our results highlight a novel strategy for modeling credit risk, and improving loan servicing mechanisms, in a scalable and privacy-preserving manner.Item Integrating Location Intelligence within Shopping Centre Reconfiguration: A Monte Carlo Simulation(2025-01-07) Wagle , Manil; Aversa, Joe; Hernandez, Tony; Doherty, SeanTechnological advancements and evolving consumer behaviors have significantly transformed the retail landscape. Traditional shopping centers, now under pressure to innovate, are leveraging location intelligence—insights from geospatial data—to stay relevant. This study explores how spatial big data can enhance leasing decisions for store expansions using Monte Carlo simulations. Focusing on a fast-fashion retailer considering expanding from 14,000 sqft to 30,000 sqft within a shopping center, it examines the impacts on sales performance. Data from Wi-Fi beacons, capturing customer traffic and sales trends from January 2022 to December 2023, provides a robust basis for analysis. The study aims to offer a quantifiable method for shopping center landlords and retailers to make informed leasing decisions under uncertainty, with two primary research objectives: leveraging spatial big data to enhance store expansion decisions and understanding how increased retail square footage impacts sales performance.Item Introduction to the Minitrack on Geospatial Big Data Analytics(2025-01-07) Aversa, Joe