Two SDSU Global Campus graduate students from the Big Data Analytics program recently attended and presented at an international research conference. 

SDSU Global Campus Big Data Graduate Pilar De Haro

We reached out to recent Big Data Analytics graduates Aichu Tan and Pilar De Haro about their experience presenting their research findings at the One Health International Conference at the Università Campus Bio-Medico di Roma in Rome, Italy. 

Give us a brief summary of the research that you presented at the conference. What motivated you to pursue this topic? 

A: “At the One Health International Conference, I presented AI for Farmers: A YOLO and Language Model-Based Plant Disease Detection and Advisory System. Our project uses computer vision to identify plant diseases from images and provides farmers with treatment and prevention advice. I was motivated by the agricultural challenges faced in many Least Developed Countries, where early disease detection can make a huge difference in crop yield, food security, and farmers’ livelihoods. My goal was to build an accessible, open-source tool that can support smallholder farmers.”

P: “At the One Health International Conference, I presented research examining how environmental, social, and systemic factors intersect to shape women’s access to healthcare along the U.S. and Mexico border. Working in collaboration with The Chiara Project and the Metabolism of Cities Living Lab, my research focused on identifying disparities in healthcare access, food security, childcare availability, socioeconomic precarity, and patterns in missing persons data across San Diego County.

I was motivated to pursue this topic not only because border regions often experience overlapping vulnerabilities that are not captured by a single dataset or discipline, but also because these issues are deeply personal to me. I am Mexican-American, and much of my adolescence was spent in San Diego while regularly crossing the border to visit family in Baja California. Growing up within this transborder context gave me firsthand insight into the complexity of border life, one that is often flattened by stereotypes or framed solely through deficit-based narratives. Women in these regions face compounded risks related to healthcare access, economic instability, and exposure to violence, yet these challenges are frequently discussed without sufficient nuance or cultural context. I wanted to use data analytics and geospatial methods intentionally, not to reinforce harmful assumptions, but to surface structural inequities with care and accuracy, and to contribute to more informed, community-centered conversations around public health and equity.”

What methods did you use when conducting your research?

A: “Our approach combined YOLOv8 for object detection to identify plant diseases and pests directly from images, alongside a language-model–based advisory system built with LangChain to generate clear and actionable treatment and prevention guidance. We collected and annotated our dataset using a crop pest and disease image collection, supplemented with extensive manual labeling through Roboflow to ensure accuracy and consistency. To make the tool widely accessible, especially for farmers in low-resource regions, we deployed the system on the web using Streamlit and Hugging Face Spaces, allowing users to easily upload images and receive instant diagnostic and advisory feedback.”

P: “I used a mixed-methods Big Data Analytics approach that combined statistical analysis with geospatial mapping. Using Python and ArcGIS Online, I cleaned, merged, and analyzed data from multiple sources, including U.S. Census variables, healthcare facility inventories, childcare provider databases, grocery access layers, and a five-year record of missing persons data from the San Diego Police Department obtained through a Public Records Act request.

I applied statistical techniques such as Pearson correlation to examine relationships between infrastructure access, gendered economic indicators, and health-related variables. I also built interactive dashboards using the ESRI ecosystem to visualize how these inequities overlap geographically, allowing the findings to be explored by both technical and non-technical audiences.”

How did you overcome any challenges that you faced during the process?

A: “Some of the biggest challenges we faced involved the quality and annotation of our dataset, as we had to manually clean and label thousands of images, which required systematic workflows and clustering techniques to manage effectively. We also had to balance accuracy with accessibility, ensuring the model remained lightweight enough to run in low-resource environments by optimizing performance and removing unnecessary components. Another challenge was integrating the detection system with the advisory component, which required multiple iterations to ensure that YOLO outputs could reliably drive meaningful LLM-generated recommendations. Ultimately, strong teamwork, continuous testing, and valuable feedback from mentors played a crucial role in helping us overcome these obstacles.”

P: “One of the main challenges of the project was working with datasets that varied widely in scale, format, and completeness. To address this, I standardized variables, normalized counts per 1,000 residents, and carefully documented all assumptions and limitations throughout the analysis. Another key challenge was interpreting correlations responsibly without overstating causality, particularly when working with sensitive social and health-related data.

Communicating complex statistical findings to non-technical audiences was an important learning process. I relied on clear visualizations, interactive dashboards, and intentional narrative framing to ensure the research remained accessible. A recognized limitation of the project was the absence of interviews with individuals or communities represented in the data, which limited the lived experiences and personal narratives that could have complemented and enriched the quantitative analysis.”

What outcomes are you hoping to achieve with your research?

A: “I hope this project contributes to improved crop health monitoring in low-resource communities and supports the development of open-source tools that student teams, researchers, and NGOs can build upon. I also aim to encourage the broader use of AI for social good, particularly in agriculture and food security. Ultimately, I want this tool to empower farmers to take early action and reduce losses caused by plant diseases and pests.”

P: “My goal is for this research to support more equitable, data-informed decision-making around public health and social services in border communities. By revealing how healthcare access, economic conditions, and safety intersect spatially, I hope the work can inform policymakers, community organizations, and researchers.

The dashboards developed through The Chiara Project are designed to remain publicly accessible and to serve as a foundation for any potential future interdisciplinary research, advocacy, and collaboration focused on women’s health and data justice.”

How has your time in SDSU Global Campus’ Big Data program helped you in your research?

A: “The Big Data Analytics program equipped me with data science, machine learning, and cloud deployment skills that were essential for this project. Courses on predictive analytics, deep learning, and decision support systems taught me how to design an end-to-end pipeline from data preprocessing to model deployment. The program also helped me build confidence in applying AI to real-world problems.”

P: “The SDSU Global Campus Big Data Analytics program played a critical role in preparing me for this research. The program strengthened my skills in data cleaning, statistical analysis, geospatial analytics, and dashboard development, while also emphasizing ethical data use and real-world application.

Equally important, the program encouraged interdisciplinary thinking and provided opportunities to present my work at professional conferences, including IEEE, One Health, and COPA. These experiences helped me refine my research questions, strengthen my communication skills, and better understand how data analytics can be used as a tool for social impact.”

What are your professional plans after graduating from SDSU?

A: “I plan to continue working in data science and machine learning, particularly on projects that integrate computer vision, cloud deployment, time series modeling, and data-driven decision systems. My recent work on dengue detection and monitoring further strengthened my interest in applying AI to public health challenges, especially through scalable cloud-based solutions that support early warning systems and strengthen community resilience. Whether in agriculture, sustainability, or health initiatives, I’m passionate about developing technology that creates meaningful social impact and improves the quality of life in underserved communities.”

P: “I plan to continue working at the intersection of data analytics and social impact. My professional goal is to apply Big Data Analytics to projects that prioritize equity, transparency, and community well-being, particularly in areas related to healthcare access, labor, policy-relevant research, data-privacy and data/AI governance.”