stenwire/NaijaFarmHand-Agent

NaijaFarmHand is an AI assistant designed specifically for beginner and early-stage farmers in Nigeria. # NaijaFarmHand: AI-Powered Farming Assistant for Nigerian Farmers **Link to Kaggle notebook(code):** kaggle.com ## Project Overview NaijaFarmHand is an AI assistant designed specifically for beginner and early-stage farmers in Nigeria. The assistant provides practical, localized, and actionable farming guidance relevant to Nigerian conditions. This project leverages Google's Gemini 1.5 Pro model through the Google AI Development Kit (ADK) to create an interactive assistant with specialized tools for Nigerian farming needs. ## Key Features - **Weather Information**: Access current weather data for specific Nigerian locations to aid in immediate farming decisions. - **Farm Planning**: Generate customized checklists for starting various farm types (crops or poultry) at different scales. - **Planting and Harvesting Guidance**: Get region-specific advice on optimal planting times and expected harvest periods for major Nigerian crops. - **Smart Research Capabilities**: Search for specific farming questions not covered by the built-in tools, including pest control, crop varieties, and market price trends. - **Nigerian Context**: All information is tailored to Nigerian farming conditions, terminology, and practices. ## Supported Farm Types - **Crops**: Maize, Cassava, Guinea Corn, Groundnut, Yam - **Livestock**: Poultry (Broiler and Layer Chickens) ## Installation Requirements This notebook requires: - An OpenWeatherMap API key (for weather data) - Google credentials configured for accessing Gemini API - Various Python packages included in the implementation ### Set up your API key The needed credential environment variable can be set to automatically configure the underlying API. This works for both the official Gemini Python SDK and for LangChain/LangGraph. To run the following cell, your API key must be stored it in a Kaggle secret with there respective names. To make the key available through Kaggle sec …