Hands-on Sessions

Introduction

Before diving into software tools and image analysis, it’s important to understand how remote sensing is used in structured, research-based investigations. This hands-on section introduces the full case study process—from designing a research question to selecting the right data, and validating your findings with field observations. 

You’ll start by learning how to design a remote sensing case study: defining the problem, choosing a study area, selecting appropriate satellite data (such as Sentinel-2), and planning analysis methods like NDVI or land cover classification. You’ll then explore how to select supporting remote sensing datasets—called ancillary data—to improve your study. These might include land use maps, administrative boundaries, road networks, or historical records, which help interpret and contextualize your results. 

Finally, you’ll focus on the importance of field (in-situ) data. Remote sensing shows you patterns from above, but field data confirms what’s actually happening on the ground. You’ll learn what types of in-situ data are useful, how to collect them (from GPS points to drone imagery), and how to integrate them into your analysis for more accurate, credible results. 

Throughout this unit, you’ll be asked to brainstorm your own case study idea, identify relevant datasets, and even step outside to collect some basic field data. This will lay the groundwork for more advanced modules using QGIS and programming tools later in the course. 

Introduction

Before working with satellite imagery or generating environmental maps, it’s essential to understand the basic types of spatial data. This short module introduces you to the two main formats used in remote sensing and GIS: vector data and raster data. You’ll explore what each format represents, how they differ, and when to use one over the other. 

Vector data—composed of points, lines, and polygons—is ideal for representing discrete features like trees, roads, or administrative boundaries. Raster data, on the other hand, stores continuous information in grid cells and is commonly used for satellite imagery, land cover, or elevation. 

You’ll also get familiar with common spatial file formats (like SHP, GeoJSON, and TIFF) and learn why metadata—the descriptive information behind every dataset—is key to making sure your data is accurate, interpretable, and useful for analysis. 

This foundational knowledge will help you better understand the structure of the datasets you’ll be working with in the hands-on parts of the course. 

Introduction

When conducting remote sensing analyses, choosing the right data source is key. In this module, you’ll get an overview of the capabilities and trade-offs of Sentinel-2, one of the most widely used Earth observation satellite systems. 

You’ll begin by understanding the core features of Sentinel-2: its spatial, spectral, and temporal resolution. Why does it matter that Sentinel-2 can revisit the same location every 5 days? Or that it captures 13 different bands of the electromagnetic spectrum? You’ll see how these characteristics make it ideal for monitoring vegetation health, land cover change, water bodies, and environmental hazards. 

At the same time, you’ll explore the limitations of this dataset. Sentinel-2, being an optical sensor, can’t penetrate cloud cover, nor can it collect data at night or provide very fine details like commercial satellites. But you’ll also learn about strategies to overcome these limits—by using multi-date images, combining datasets (e.g., with radar or thermal sources), or applying cloud-based processing tools like Google Earth Engine or Sentinel Hub. 

This module prepares you to make informed decisions about when and how to use Sentinel-2, and how to complement it with other spatial data to get the clearest and most accurate insights possible. 

Introduction

Spatial data analysis is one of the most powerful tools for understanding and managing the world around us. In this module, you’ll explore both the strengths and limitations of working with geospatial information. From tracking urban expansion and environmental degradation to mapping disease risk or flood zones, spatial data helps reveal patterns that are invisible in traditional tabular datasets. 

You’ll learn how spatial analysis works—combining location (the “where”) with attributes (the “what,” “when,” and “why”) to answer real-world questions. The examples you’ll see cover everything from wetland monitoring to urban planning and agricultural management, highlighting how geodata supports decision-making and real-time environmental monitoring. 

But geodata isn’t perfect. You’ll also examine common challenges, including cloud cover in satellite images, resolution trade-offs, limited access to high-quality datasets, and the technical demands of processing large files. Ethical considerations—like data privacy and ownership—also come into play. 

Fortunately, there are solutions: using open-access platforms like Google Earth Engine, combining multiple datasets for better accuracy, and applying FAIR data principles to ensure responsible and reproducible use. 

This module encourages critical thinking: spatial data is powerful, but only if used wisely and with a clear understanding of its strengths and constraints. 

Introduction

This module marks the real start of the practical part of the Remote Sensing Fundamentals course. From here on, you’ll not only learn about spatial data—you’ll start working with it directly, using two powerful and widely used open-source tools: QGIS and RStudio. 

You’ll begin by understanding what spatial software is, why it matters, and when to choose GIS platforms, cloud-based tools, or coding environments. Then, you’ll dive into QGIS, learning how to explore, visualize, and manage spatial datasets through an intuitive interface. Topics include loading shapefiles, adjusting coordinate reference systems (CRS), installing essential plugins, and saving project files. 

Next, you’ll get introduced to R and RStudio—a programming environment ideal for statistical and spatial analysis. You’ll learn how to navigate the RStudio interface, create and assign variables, work with data types and libraries, and write simple scripts for geospatial tasks. 

Alongside the slides, you’ll receive ZIP folders containing hands-on exercises, sample scripts, and small assignments. These tasks will help reinforce your learning, offering step-by-step guidance while also encouraging experimentation. 

By the end of this module, you’ll be fully set up and confident using both QGIS and R—ready to analyze real geospatial data in the modules to come. 

Introduction

Now that you’ve set up your spatial software, it’s time to start working directly with real geospatial datasets. In this module, you’ll learn how to import, inspect, manipulate, and visualize two essential types of spatial data: vector data (points, lines, polygons) and raster data (gridded image layers like elevation or satellite imagery). 

Using both QGIS and RStudio, you’ll practice: 

• Loading shapefiles and GeoTIFFs 

• Exploring coordinate systems and metadata 

• Visualising data with different band combinations (e.g., false-colour imagery) 

• Running basic spatial operations like masking, resampling, and reprojecting 

The exercises are designed to build your fluency in handling data across both platforms—leveraging QGIS for intuitive, visual tasks and R for repeatable, script-based workflows. Along the way, you’ll gain practical insight into how spatial layers behave, how they’re structured, and how you can start building your own spatial analyses step by step. 

Introduction

In this module, you’ll explore how vector-based spatial analysis can answer real-world questions using tools built into QGIS and R. Vector data—points, lines, and polygons—is often used to represent infrastructure, natural features, and administrative boundaries. You’ll use this data to identify spatial relationships, assess proximity and overlap, and communicate your findings with clear, professional maps. 

The hands-on exercises centre on a practical case: evaluating how road infrastructure affects wetlands in Botswana. You’ll learn to: 

• Use OpenStreetMap (OSM) data to extract roads and wetlands 

• Clip data to a defined area of interest (AOI) 

• Reproject layers for accurate measurement 

• Apply buffer analysis to model zones of influence around wetlands 

• Perform intersection analysis to find roads within critical distances 

• Visualize and export your results in a well-designed map layout 

Whether it’s for environmental impact assessments, land-use planning, or conservation mapping, vector analysis forms the core of many spatial workflows. This module will train you not only in technical operations, but also in making data-driven decisions and presenting your results clearly and convincingly. 

Introduction

This module introduces you to essential raster preprocessing techniques—steps that transform raw satellite imagery into clean, analysis-ready data. When working with optical sensors like Sentinel-2, challenges such as cloud cover, fragmented image tiles, and varying resolutions can complicate your work. Learning how to handle these limitations is a core skill in remote sensing. 

You’ll start by learning how to acquire Sentinel-2 data using the Copernicus browser, including how to define your area of interest, filter scenes by cloud cover, and download imagery in a format suitable for QGIS. From there, you’ll dive into cloud masking—removing unwanted pixels using quality bands or automated tools—and practice mosaicking, the process of stitching multiple images into a single, continuous raster. 

Using both standard QGIS tools and the Semi-Automatic Classification Plugin (SCP), you’ll go through real exercises that guide you step by step: from inspecting band properties, to building virtual rasters, to generating true-color or false-color composites. You’ll also explore how to apply masks, interpret cloud classification layers, and troubleshoot visual artifacts caused by overlapping tiles or NoData values. 

By the end of this module, you’ll understand the workflow needed to clean and prepare raster datasets, giving you a solid foundation for the more advanced analysis ahead. 

Introduction

In this comprehensive module, you will dive deep into raster analysis, one of the most critical and data-rich parts of remote sensing. You’ll learn how to extract meaningful information from satellite imagery by using mathematical indices and machine learning classification techniques—combining spatial analysis with ecological understanding. 

We begin with vegetation indices, learning how reflectance in different spectral bands can reveal vegetation health, water content, or moisture stress. You’ll calculate NDVI, NDWI, and NDMI using Sentinel-2 imagery, and understand when and why to use each index in different environmental contexts. These indices are more than just numbers—they are diagnostic tools used in agriculture, conservation, climate studies, and land degradation monitoring. 

Building on that, the module moves into land cover classification. You’ll explore both supervised and unsupervised approaches, comparing popular algorithms like Maximum Likelihood and Random Forest. Then, you’ll put theory into practice by developing your own land cover classification project using QGIS for training data creation and RStudio for model training and validation. Along the way, you’ll encounter real-world challenges like class imbalance, spectral confusion (e.g., wetlands vs. water), and the importance of high-quality training data. 

This module blends conceptual understanding with hands-on workflows—guiding you through everything from raster calculations to producing high-resolution land cover maps. Whether you’re interested in forest monitoring, wetland mapping, or urban expansion, this section equips you with essential skills to translate raw imagery into actionable, spatial intelligence. 

Introduction

After completing a land cover classification, it’s not enough to produce a map—you need to verify that your results are reliable. This module introduces you to validation and accuracy assessment, the final and essential step in the remote sensing analysis pipeline. 

You’ll explore the concept of the confusion matrix, a tool that compares your classified data to known reference data. This matrix forms the basis for calculating key metrics such as Overall Accuracy, Producer’s Accuracy, and User’s Accuracy—each revealing different aspects of classification performance. You’ll learn how to interpret these values, identify where misclassifications occurred, and understand the common causes of error, such as sensor noise, poor training data, or spectral overlap between classes. 

The module guides you through practical validation workflows using both QGIS (via the Semi-Automatic Classification Plugin) and RStudio, showing how to input validation data, run assessments, and generate interpretable results. Working with a real classification from the previous module, you’ll apply stratified sampling, calculate accuracy scores, and visualize where your model succeeded—or fell short. 

Ultimately, accuracy assessment helps you build confidence and transparency in your results. It’s not just a technical step—it’s a way of ensuring your spatial outputs can be trusted by researchers, decision-makers, and communities who rely on the insights you generate. 

Introduction

This final module is a hands-on challenge designed to help you apply everything you’ve learned throughout the course. You’ll design and carry out a complete remote sensing case study, from data selection and preprocessing to analysis, validation, and communication of results. 

Using both vector and raster data, you’ll define your own research question—something relevant to land cover, water bodies, vegetation change, or any other theme you’ve explored. You’ll choose appropriate remote sensing and ancillary datasets, perform preprocessing and analysis steps (such as classification or vegetation index calculation), and assess the accuracy of your results. 

This task is also an exercise in scientific reporting. You’ll be guided through how to write a short report or presentation that explains your methodology, visualizes your findings, and reflects on the outcomes. You’ll need to demonstrate your understanding of spatial resolution, spectral bands, QGIS operations, R-based scripting, validation metrics, and data interpretation. 

Think of this as your first real remote sensing project—a small-scale but complete workflow that mirrors what professionals and researchers do in the field. It’s not just a test; it’s a way to see how much you’ve grown, and to begin shaping the kinds of spatial stories you want to tell with your own data.