P Omar R T F O L I O

Problem:Farmers lack fast, affordable plant-disease diagnosis → delayed treatment and crop loss.

Built:CNN/ResNet deep-learning model classifying 186 plant disease types from leaf images + a treatment-recommendation engine using a Kaggle cures dataset. One leaf photo → diagnosis + actionable treatment.

Built:Collected & preprocessed a Kaggle plant-disease image dataset; trained/evaluated a classifier with data augmentation and hyperparameter tuning; became the core model of the AgriTech project.

Problem:Visibility into 50,000+ orders across regions, categories and customer segments.

Built:Cleaned data in Excel Power Query, built an interactive Power BI dashboard (revenue by region, top categories, monthly trends, profit margins) + stakeholder-ready PowerPoint with recommendations.

Problem:Understand what drives turnover across 1,400+ employee records.

Built:EDA on age, department, job satisfaction, salary, work-life balance; visualized attrition by department, gender, tenure (Power BI + Excel pivots); delivered a retention-strategy PowerPoint.