UC GIS Week

ElephantTracker: Spatiotemporal Analytics for Human–Elephant Conflict Mitigation

https://elephanttracker.com/

The ElephantTracker website is a research and conservation platform designed to help analyze and visualize elephant movement and human-elephant conflict patterns.It provides interactive GIS maps to display historical GPS tracking data, showing elephant movement paths, coordinates, and timelines. Researchers and conservationists can upload and manage datasets, making it easier to centralize and analyze elephant tracking information.The site also links elephant movements with records of human-elephant conflict incidents, such as crop damage or property loss. By doing so, it helps identify patterns and geographic hotspots where conflict is more likely to occur. These insights can support risk assessment, allowing conservation teams and local authorities to develop targeted strategies to reduce conflict and protect both communities and elephants.In short, ElephantTracker’s purpose is to serve as a decision-support tool for researchers, NGOs, and wildlife managers by providing clear, data-driven insights into elephant movement and its relationship with human activity. This helps guide conservation priorities, improve coexistence strategies, and ultimately support the long-term survival of elephant population.

Mapping Californian parking lots using machine learning to identify suitable for large-scale solar development

Parking lots represent a currently underutilized potential for solar in the built environment. Currently, there is no statewide map of parking lots in California, making it difficult to prioritize the best sites for solar installation. This project uses machine learning to identify and classify parking lots that may be suitable for solar development. First, we use parcel, roadway, railway, and other land use data to identify suitable parcels for large-scale solar. Next, we classified NAIP infrared imagery in these suitable parcels using machine learning in the ArcGis Pro Imagery classification system. We then post-processed to improve accuracy and calculate vegetation coverage in each parking lot. This map will not only be useful for solar development but can be used in further research in a wide variety of areas, including urban tree cover, land use and planning, and rainwater infiltration.

Uneven Validity of Place — Spatial Heterogeneity in POI Temporal Validity

Point-of-interest (POI) datasets are increasingly used in geographic research to study urban density, accessibility, and spatial clustering. Yet, these datasets often assume that each POI accurately reflects real-world conditions at the time of analysis. This presentation investigates the temporal validity of POIs—whether a place recorded in digital maps still exists and operates at the time of data use. Focusing on Amap POI data from Guangzhou, China (2014–2024), I evaluate temporal validity using both intrinsic and extrinsic approaches. Intrinsic measures assess the lag between the dataset’s retrieval date and the POI’s recorded update time, revealing spatial heterogeneity linked to land use, population, and POI diversity. Extrinsic validation compares POIs with historical street view imagery to identify outdated or demolished places. Results demonstrate uneven temporal accuracy across the urban landscape, emphasizing the need to quantify and model bias in volunteered geographic data. By assigning confidence scores based on update time and persistence, this work contributes to more reliable use of POI data in urban research and planning.