Crop Disease Outbreak Forecasting With Geospatial Risk Maps

Added Jun 2025 3 design docs

Crop diseases move faster than manual scouting: by the time blight is visible across a field, the treatment window has often closed. Forecasting outbreaks from weather, soil, and sensor data gives growers days or weeks of warning, but it requires processing volumes of geospatial data that overwhelm desktop tools. The intern builds a forecasting platform on a distributed stack. Hadoop provides the storage backbone for large historical datasets, Spark runs the heavy processing that joins weather, soil, and field sensor data across regions, and pandas handles feature engineering and analysis in Python. Predictive models estimate outbreak risk by crop and location, and the results render as geospatial risk maps with early-warning alerts. Users upload their own datasets, visualize risk across their fields, and receive alerts as conditions shift, while Docker keeps the multi-service platform reproducible from development through deployment. The project delivers real distributed data engineering experience, with Hadoop and Spark working together on a problem where scale is unavoidable, plus geospatial analytics and the responsibility of communicating probabilistic forecasts to people whose livelihoods depend on getting them right.

Related projects

Crop Disease Classification Service Using Transfer Learning

Crop diseases are often identified too late for effective intervention because farmers and field agronomists rarely have quick access to plant pathology expertise. A photograph of a diseased leaf contains enough signal for accurate diagnosis, but turning that signal into a dependable, scalable service is a genuine engineering challenge for agricultural research groups. The intern builds a full diagnostic platform around transfer learning: pre-trained ResNet and MobileNet backbones are fine-tuned in PyTorch on crop disease imagery, with OpenCV and NumPy handling preprocessing and a data augmentation pipeline improving robustness to lighting and orientation. A Streamlit interface lets users upload leaf photos and receive ranked disease predictions alongside explainable AI visualizations that highlight the image regions driving each classification, while a FastAPI service exposes a batch inference endpoint for research workloads and pandas powers summary reporting. The entire system is containerized with Docker so it can be deployed consistently across lab and cloud environments. The project demonstrates end-to-end applied computer vision: adapting pre-trained networks to a specialist domain, communicating model reasoning to non-technical users, and packaging research code as a deployable, reproducible service.

agricultureeducationcomputer-visiondocker+9

View project

Medium

Wildlife Species Identification with Fine-Tuned ResNet Models

Conservation researchers collect enormous volumes of wildlife imagery from camera traps and field surveys, but manually identifying the species in every frame is a bottleneck that delays population studies and habitat decisions. Automated species classification lets small research teams process seasons of imagery in hours instead of months. The intern builds a Streamlit application centered on transfer learning: a ResNet model pre-trained on general imagery is fine-tuned in PyTorch to classify wildlife photos into multiple species. The project includes a data augmentation pipeline using OpenCV and NumPy to simulate the messy conditions of field photography, controls for staged fine-tuning of the network layers, and a prediction dashboard that reports ranked species probabilities for each uploaded image. An explanation view highlights the image regions most responsible for each prediction, helping ecologists confirm the model is keying on the animal rather than the background, while pandas tracks evaluation metrics across training runs. The intern comes away understanding how to adapt large pre-trained vision models to a specialized dataset, how augmentation and fine-tuning choices affect generalization, and how to present model evidence in a form that earns the trust of scientific users.

agricultureresearchcomputer-visionnumpy+5

View project

Hard

Smart Farm Data Platform with Delta Live Tables Orchestration

Modern farms generate relentless data, soil sensors, weather stations, and drone imagery telemetry, but agronomy teams rarely have engineering support to turn those feeds into trustworthy analytics. The gap is architectural: raw sensor streams need ingestion, validation, transformation, and governance before anyone can responsibly base irrigation or planting decisions on them. The intern designs an end-to-end smart agriculture platform on Databricks and Spark. Sensor and drone data arrive through Kafka streams and Autoloader file ingestion, then flow through a Medallion Architecture implemented with Delta Live Tables: Bronze tables capture raw records, Silver tables apply DLT expectations that quarantine bad readings, and Gold tables aggregate field-level metrics. Databricks Workflows and Airflow orchestrate the pipelines on schedules and dependencies, while Databricks SQL dashboards visualize crop health indicators and yield predictions from the curated layer. The intern applies Unity Catalog for governance, secret scopes for credential handling, and cluster policies aimed at cost control. The project demonstrates command of the modern lakehouse stack, streaming plus batch ingestion, declarative pipelines with enforced data quality, dual orchestration tools, and security practices, applied to a domain where data quality has physical consequences.

agricultureairflowbig-data+11

View project