Introduction

Welcome to our Geospatial Science learning project! We have created this guide that will help you set up your learning environments, create heatmaps using QGIS and Python, and visualize sorghum research field trial data while developing data science and machine learning skills. We’ll walk you through how to build a predictive model for important plant traits (e.g., yield, seed protein content, and chlorophyll a fluorescence) using data from drones, field instruments, technicians, and climate records.

You’ll also learn to track and interpret key metrics—such as preprocessing steps, hyperparameter tuning, and feature importance—by installing and using TensorBoard. By following these instructions, you’ll be prepared to analyze geospatial data and apply machine learning techniques to understand plant performance more effectively.

What You Will Learn

1. Geospatial Concepts: Understand the fundamentals of geospatial data, GIS systems, and their applications.
2. Data Preparation: Learn how to clean, preprocess, and structure data for analysis.
3. Modeling and Predictions: Implement and evaluate regression models to predict plant traits like yield or protein content.
4. Tool Proficiency: Gain hands-on experience with tools like JupyterLab, Google Colab, and programming libraries in Python.

Prerequisites

Before starting this project, ensure you have access to the following:

Tools and Software

• Python (3.9 or higher): Programming language for coding the project.
• Visual Studio Code: This is a popular IDE, or Integrated Development Environment, used for running code.
• Google Colab or JupyterLab: Platforms for running Python scripts and notebooks.
• Excel or CSV Editor: To view and manage data files.

Required Libraries

You will need the following Python libraries:

• pandas
• numpy
• scikit-learn
• matplotlib
• openpyxl

If you are unsure how to install these, don’t worry. Detailed setup instructions are provided in the specific sections for JupyterLab and Google Colab.

Project Overview

This project focuses on:

• Setting up your environment for data analysis, creating heatmaps with QGIS and Python, and visualizing sorghum field trial data.
• Developing a predictive model to estimate plant traits—such as yield, protein content, or chlorophyll fluorescence—using data from drones, field instruments, technicians, and climate records.
• Installing and using TensorBoard to log, track, and visualize various metrics (e.g., preprocessing steps, hyperparameter tuning, and feature importance) throughout the modeling process.

Data Sources

You will work with:

• Wireless sensor data
• Climate data
• Field technician observations
• Crop data

Models

The project includes six pre-implemented regression models:

• Partial Least Squares Regression (PLS)
• Lasso Regression
• ElasticNet Regression
• Bayesian Ridge Regression
• Extreme Gradient Boosting (XGBoost)
• Support Vector Regression (SVR)

You are encouraged to explore these models and even implement your own, such as Random Forest or Linear Regression.

Quick Start

Already familiar with development environments like Google Colab and JupyterLab? Great! Use the exercises below to refresh your skills and get hands-on practice with these tools. These mini-tasks will help you set up your environment and ensure everything is ready for the project.

Google Colab

1. Open Google Colab.
2. Install required libraries using:

   !pip install pandas numpy scikit-learn matplotlib openpyxl
   

3. Upload your dataset and start coding!

JupyterLab

1. Install JupyterLab using:

   pip install jupyterlab
   

2. Set up your environment by installing the required libraries:

   pip install pandas numpy scikit-learn matplotlib openpyxl
   

3. Open JupyterLab and create a new notebook for the project.

Tips for Success

• Read the Instructions: Carefully follow the guides for setting up and running the code.
• Stay Organized: Keep your data, scripts, and results in a well-structured directory.
• Experiment: Don’t hesitate to tweak model parameters and explore alternative approaches.
• Ask for Help: Reach out to your instructor or peers if you encounter challenges.

Getting Help

If you encounter challenges or have questions, there are several resources to support you:

• Ask Your Instructor: Your teacher is there to help clarify instructions and troubleshoot issues.
• Google It: Use search engines to look up terms or code that you find challenging.
• Review Documentation: Refer to this guide or library-specific documentation for deeper insights.

Submitting an Issue

If you find an issue with the project or need further assistance, feel free to reach out to the lab directly:

• Email: shakoorlab.danforth@gmail.com

We appreciate your feedback and will do our best to assist you promptly!