Joshua Pitts’s research while affiliated with Boston University and other places

This study presents a comparative analysis of two leading large language models (LLMs), Google’s Bard, powered by the Gemini model, and OpenAI’s ChatGPT 4.0, in the context of their responses to coastal ecosystem science undergraduate student education. Fifty questions related to coastal ecosystem management were posed to each LLM. Expert assessments evaluated the responses of the LLMs based on five key metrics: accuracy, relevance, depth, creativity, and semantic clarity. Knowledge graphs provided a structured framework for assessing and visualizing the AI responses. The analysis identified the strengths and weaknesses of each LLM in addressing complex environmental issues. The findings contribute to a deeper understanding of LLMs’ potential applications in environmental science education and scientific communication. This study acknowledges limitations, such as the inherent subjectivity of expert assessments and the potential for bias within the knowledge graphs used for evaluation. Future research directions include investigating the effectiveness of LLMs in personalized learning environments and exploring their potential for generating educational content tailored to diverse audiences.

The semiconductor industry is essential to modern global economies, as chips and other components are crucial in consumer and industrial goods. In the past decade, the explosion of new technologies, such as the Internet of Things (IoT), Big Data, Artificial Intelligence (AI), and 5G telecommunication infrastructure has created a sustained demand for semiconductors that is reshaping every industry on its path to digitization and automation. We first characterize the supply chain of semiconductors, which over time have been honed to deliver maximum efficiency and speed, and we examine the drivers of past disruptions, especially due to the COVID-19 pandemic, natural hazards, and increasing geopolitical tensions between China and the West. Second, we present the rationale for resiliency management in this critical sector by probing what countries and companies plan to do given these disruptions. Finally, we are examining the role of Geographic Information Systems (GIS), spatial analysis, and AI in semiconductors supply chain management. While the specific tools and analytics for supply chain analysis remain an open question for researchers and managers, in order to model, predict, and plan for larger scale disruptions, such as pandemics, they definitely will rely heavily on data-driven approaches.

China is now the world’s largest source of bilateral development finance and will likely continue to play a prominent role in sovereign lending through its multi-billion-dollar Belt and Road Initiative. This paper introduces major methodological enhancements in tracking this finance: the use of an original application programming interface (API) to gathers news in multiple languages; double-verification of every record to ensure every finance commitment has been formalized; and visual geo-location to trace the precise footprint of every project. The resulting dataset enables economic, environmental, and social analyses with high-precision spatial accuracy, as well as spatiotemporal monitoring by project stakeholders and enhanced planning by project managers. It covers the years 2008–2019 to enable analysis before and after the announcement of the Belt and Road Initiative. It includes 862 finance commitments, 669 of which have geographic location, to 94 countries across the world.

With the world population projected to grow significantly over the next few decades, and in the presence of additional stress caused by climate change and urbanization, securing the essential resources of food, energy, and water is one of the most pressing challenges that the world faces today. There is an increasing priority placed by the United Nations (UN) and US federal agencies on efforts to ensure the security of these critical resources, understand their interactions, and address common underlying challenges. At the heart of the technological challenge is data science applied to environmental data. The aim of this special publication is the focus on big data science for food, energy, and water systems (FEWSs). We describe a research methodology to frame in the FEWS context, including decision tools to aid policy makers and non-governmental organizations (NGOs) to tackle specific UN Sustainable Development Goals (SDGs). Through this exercise, we aim to improve the “supply chain” of FEWS research, from gathering and analyzing data to decision tools supporting policy makers in addressing FEWS issues in specific contexts. We discuss prior research in each of the segments to highlight shortcomings as well as future research directions.

The United Nations' Sustainable Development Goal 3 is to ensure health and well-being for all at all ages with a specific target to end malaria by 2030. Aligned with this goal, the primary objective of this study is to determine the effectiveness of utilizing local spatial variations to uncover the statistical relationships between malaria incidence rate and environmental and behavioral factors across the counties of Kenya. Two data sources are used-Kenya Demographic and Health Surveys of 2000, 2005, 2010, and 2015, and the national Malaria Indicator Survey of 2015. The spatial analysis shows clustering of counties with high malaria incidence rate, or hot spots, in the Lake Victoria region and the east coastal area around Mombasa; there are significant clusters of counties with low incidence rate, or cold spot areas in Nairobi. We apply an analysis technique, geographically weighted regression, that helps to better model how environmental and social determinants are related to malaria incidence rate while accounting for the confounding effects of spatial non-stationarity. Some general patterns persist over the four years of observation. We establish that variables including rainfall, proximity to water, vegetation, and population density, show differential impacts on the incidence of malaria in Kenya. The El-Nino-southern oscillation (ENSO) event in 2015 was significant in driving up malaria in the southern region of Lake Victoria compared with prior time-periods. The applied spatial multivariate clustering analysis indicates the significance of social and behavioral survey responses. This study can help build a better spatially explicit predictive model for malaria in Kenya capturing the role and spatial distribution of environmental, social, behavioral, and other characteristics of the households.

Global financial investments in energy production and consumption are significant since all aspects of a country’s economic activity and development require energy resources. In this paper, we assess the investment trends in the global energy sector during, before, and after the financial crisis of 2008 using two data sources: (1) The Dealogic database providing cross-border mergers and acquisitions (M&As); and (2) The “fDi Intelligence fDi Markets” database providing Greenfield (GF) foreign direct investments (FDIs). We highlight the changing role of China and compare its M&A and GF FDI activities to those of the United States, Germany, UK, Japan, and others during this period. We analyze the investments along each segment of the energy supply chain of these countries to highlight the geographical origin and destination, sectoral distribution, and cross-border M&As and GF FDI activities. Our paper shows that while energy accounts for nearly 25% of all GF FDI, it only accounts for 4.82% of total M&A FDI activity in the period 1996–2016. China’s outbound FDI in the energy sector started its ascent around the time of the global recession and accelerated in the post-recession phase. In the energy sector, China’s outbound cross-border M&As are similar to the USA or UK, located mostly in the developed countries of the West, while their outbound GF investments are spread across many countries around the world. Also, China’s outbound energy M&As are concentrated in certain segments of the energy supply chain (extraction, and electricity generation) while their GF FDI covers other segments (electricity generation and power/pipeline transmission) of the energy supply chain.

Global financial investments in energy production and consumption are significant since all aspects of a country's economic activity, and development require energy resources. In this paper, we assess the investment trends in the global energy sector during, before and after financial crises of 2008 using two data sources: (1) Dealogic database providing cross‐border mergers and acquisitions (M&As), and (2) fDi Intelligence fDi Markets database providing greenfield (GF) foreign direct investments (FDIs). We highlight the changing role of China and compare its M&A and GF FDI activities to those of the United States, Germany, UK, Japan and others during this period. We analyze the investments along each segment of the energy supply chain of these countries to highlight the geographical origin and destination, sectoral distribution, and cross‐border M&As and GF FDI activities. Our paper shows that while energy accounts for nearly 25% of all GF FDI, it only accounts for 4.82% of total M&A FDI activity in the period 1996-2016. China's outbound FDI in the energy sector started its ascent around the time of the global recession and had accelerated in the post-recession phase. In the energy sector, the development of China's outbound cross‐border M&As is similar to USA or UK, located mostly in the developed countries in the west, while their outbound GF investments are spread across many countries around the world. Also, China's outbound energy M&As are concentrated in certain segments (extraction, and electricity generation) while their GF covers all segments of the energy supply chain.

Characterizing urban landscapes is important given the present and future projections of global population that favor urban growth. The definition of “urban” on a thematic map has proven to be problematic since urban areas are heterogeneous in terms of land use and land cover. Further, certain urban classes are inherently imprecise due to the difficulty in integrating various social and environmental inputs into a precise definition. Social components often include demographic patterns, transportation, building type and density while ecological components include soils, elevation, hydrology, climate, vegetation and tree cover. In this paper, we adopt a coupled human and natural system (CHANS) integrated scientific framework for characterizing urban landscapes. We implement the framework by adopting a fuzzy sets concept of “urban characterization” since fuzzy sets relate to classes of object with imprecise boundaries in which membership is a matter of degree. For dynamic mapping applications, user-defined classification schemes involving rules combining different social and ecological inputs can lead to a degree of quantification in class labeling varying from “highly urban” to “least urban”. A socio-economic perspective of urban may include threshold values for population and road network density while a more ecological perspective of urban may utilize the ratio of natural versus built area and percent forest cover. Threshold values are defined to derive the fuzzy rules of membership, in each case, and various combinations of rules offer a greater flexibility to characterize the many facets of the urban landscape. We illustrate the flexibility and utility of this fuzzy inference approach called the Fuzzy Urban Index for the Boston Metro region with five inputs and eighteen rules. The resulting classification map shows levels of fuzzy membership ranging from highly urban to least urban or rural in the Boston study region. We validate our approach using two experts assessing accuracy of the resulting fuzzy urban map. We discuss how our approach can be applied in other urban contexts with newly emerging descriptors of urban sustainability, urban ecology and urban metabolism.