2 Chapter 2: Making Sense of Spatial Data Using GIS & Geocoding and Georeferencing

2.1 Introduction

“Everything is related to everything else, but near things are more related than distant things.” – Waldo Tobler’s First Law of Geography.

This simple yet profound law forms the foundation of geospatial analysis. When we model and analyze the world, location is not just a variable; it’s a critical dimension that influences how things interact. Spatial proximity often drives relationships that might be missed in non-spatial models. For instance, predicting housing prices based purely on characteristics like size or age can miss the significant impact of location. A house next to a park or near public transportation will often command a higher price, even if it’s identical to one farther away.

In this way, geospatial data allows us to enhance traditional models by incorporating the spatial dimension—a layer of truth that non-spatial models overlook.

Geospatial data science sits at the crossroads of geography and data science, providing a framework for understanding the world through the lens of location. In today’s data-driven society, location matters more than ever. From smartphones equipped with GPS to vast networks of sensors, the explosion of geospatial data has revolutionized how we analyze global phenomena—whether it’s tracking climate change, analyzing economic trends, or mapping global migration patterns.

This chapter will guide you through the foundational concepts and tools used in geospatial data science, focusing on how spatial data is represented, managed, and analyzed, primarily through the use of QGIS and R. It provides a comprehensive overview of essential tools available for geospatial data analysis, with a primary focus on QGIS and R. As geospatial data becomes increasingly prevalent across various fields—ranging from aerial photography and satellite imagery to census data and geolocated social media posts—understanding how to process and analyze this data is crucial.

We will explore the two primary types of geospatial data:

Raster data: Represents variables at each point in a grid covering the data’s full extent (e.g., satellite images).
Vector data: Associates variables with discrete geometric objects in space (e.g., city locations on a map).

While QGIS offers a user-friendly graphical interface that simplifies many tasks, script-based processing in R remains essential for certain operations due to its reproducibility, transparency, and integration with other datasets. Throughout this chapter, we will demonstrate how to effectively use QGIS for various geospatial operations, with R code examples included for those interested in script-based analysis.

2.2 The Prevalence of Geospatial Data

Today, geospatial data is ubiquitous. From satellites capturing high-resolution images of the Earth to the growing availability of government data, organizations and researchers now have access to vast amounts of spatial data.

The power of spatial data lies in its ability to show how things relate to one another geographically. For instance, analyzing the spatial distribution of economic activity alongside transportation networks can reveal how access to infrastructure impacts growth.

Concepts

This section offers an overview of the tools available in R for analyzing geolocated data. This type of data is becoming increasingly common in various domains, such as aerial photography, satellite imagery, census data, and geolocated social media posts.

Geospatial data generally falls into two categories:

Raster data: Represents variables defined at each point of a grid covering the full extent of the data (e.g., satellite images).
Vector data: Associates variables (sometimes called attributes) with discrete geometric objects located in space (e.g., city locations on a map).

Objectives of Script-Based Analysis

While the graphical interface of GIS software like QGIS is intuitive, script-based processing in R offers several advantages:

Reproducibility: It is easy to repeat the analysis for new data by rerunning the script.
Transparency: Other researchers can reproduce the methods if they have access to the same programming language.
Integration: Spatial data can be extracted and merged with other datasets for statistical analysis within a single programming environment.

Goals

Familiarize yourself with R packages for processing and visualizing vector and raster data (e.g., sf and stars).
Perform common data transformation operations using these packages.
Create more complex static maps (using ggplot2) and interactive maps (using mapview).

Note on Packages

The set of packages available for spatial analysis in R has rapidly evolved. A few years ago, the sp and raster packages were the main tools for processing vector and raster data, respectively. sf and stars are part of a recent initiative to overhaul R’s spatial tools (https://www.r-spatial.org/).
The sf package represents spatial data frames with a standard format based on open-source geodatabases and integrates well with popular R packages for data manipulation and visualization (such as dplyr and ggplot2).
The stars package is also compatible with ggplot2 and provides good support for raster “cubes” with non-spatial dimensions such as time.
The raster package and its successor terra (released in 2020) possess some features not present in stars and perform some operations more quickly. Therefore, it may be useful to learn them if your workflow involves complex operations on large rasters; see the documentation on https://r-spatial.org/ for more details.

2.3 Working with Geospatial Data in QGIS and R

Geospatial data science combines the principles of geography with data science to analyze phenomena through the dimension of location. In this chapter, we’ll explore the foundational concepts and tools used in geospatial data science, with a focus on how spatial data is represented, analyzed, and visualized using QGIS and R.

Why Location Matters

Waldo Tobler’s First Law of Geography states: “Everything is related to everything else, but near things are more related than distant things.” Location is a fundamental aspect of how the world functions, and by incorporating the spatial dimension, we can better understand relationships that may be overlooked in non-spatial models. This concept is central to geospatial analysis, which adds significant value to predictions, such as the relationship between housing prices and their proximity to infrastructure or amenities.

Types of Geospatial Data

Geospatial data comes in many forms, each suited for different types of analysis. Here are the most common data formats and types:

Data Formats:

Data Type	Non-Spatial Formats	Spatial Formats
Text	CSV, JSON, XML	GeoJSON, GML, KML
Binary	PDF, XLS, ZIP	Shapefile, GeoPackage
Images	TIFF, JPG, PNG	GeoTIFF, JPEG2000
Databases	SQLite, PostgreSQL	Spatialite, PostGIS

Shapefiles: Widely used for storing geometries and attributes.
GeoJSON: Lightweight and web-friendly format for spatial data.
GeoTIFF: Common for raster data, such as satellite images.

Data Types:

Vector Data: Points, lines, and polygons representing features like roads, cities, or administrative boundaries.
Raster Data: Gridded data that represents variables such as temperature or elevation, widely used in remote sensing.
Tiles: Subdivided raster or vector datasets, commonly used in web mapping.

Understanding Spatial Data: Geometry and Attributes

Spatial data is composed of two critical elements:

Geometry: This describes the physical shape and location of spatial features (points, lines, polygons).
Attributes: Additional data describing the features, such as population, area, or economic activity.

Example: GeoJSON Format

A GeoJSON object represents spatial data by encoding geometry (coordinates) and attributes. Here is a simple GeoJSON for the city of Montreal:

{
  "type": "Feature",
  "geometry": {
    "type": "Point",
    "coordinates": [-73.5673, 45.5017]
  },
  "properties": {
    "id": 1,
    "name": "Montreal"
  }
}

This example includes both the geometry (longitude and latitude) and attributes (city name and ID).

Coordinate Reference Systems (CRS) and Map Projections

CRS are essential for representing the Earth’s 3D surface on 2D maps. Choosing the right CRS ensures accuracy in spatial analysis.

Map Projections:

Equal Earth projection: Ideal for global maps, preserving area proportions.
UTM (Universal Transverse Mercator): Commonly used for regional mapping due to minimized distortions in specific zones.

Example: Choosing a CRS

For global maps, projections like Equal Earth offer visual appeal and minimal distortion. For local studies, country-specific CRSs, such as EPSG:7755 for India, are used to maintain accuracy over small areas.

Tools for Geospatial Data Science: QGIS and R

QGIS: A Practical Introduction

QGIS is an open-source GIS platform that offers a wide range of tools for visualizing, manipulating, and analyzing geospatial data. Its ability to handle multiple data formats, overlay layers, and visualize complex relationships makes it ideal for both beginners and advanced users.

Key QGIS functions include: - Data loading: Use the Data Source Manager to load shapefiles, raster data, and other formats. - Exploring attributes: Tools like the Identify tool allow you to view and explore attribute data. - Applying symbology: Customize how data is visualized using the Layer Styling Panel to create meaningful maps.

R for Geospatial Analysis

R complements QGIS by offering advanced statistical capabilities. With packages like sf and ggplot2, R provides powerful tools for spatial analysis and data visualization. For example, the sf package allows seamless handling of spatial data, while ggplot2 enables sophisticated data visualization.

Example: Loading and Exploring Data in R

You can perform operations similar to QGIS using the sf package in R:

library(sf)
mrc <- st_read("data/mrc.shp")
st_bbox(mrc)
st_crs(mrc)
head(mrc)

Working with Geospatial Data: Visualization and Manipulation

Visualizing Geospatial Data in QGIS

In QGIS, you can visualize data by: 1. Loading data: Use the Data Source Manager to import data. 2. Apply Symbology: Open the Layer Styling Panel to customize the visualization based on attributes.

In R, you can use ggplot2 to create maps:

ggplot(data = mrc_proj) +
  geom_sf()

Manipulating Geospatial Data in QGIS

QGIS offers various tools for data manipulation: - Field Calculator: Perform calculations or create new fields. - Processing Toolbox: Tools like buffering, clipping, and spatial joins.

In R, similar data manipulation can be performed using dplyr and sf:

regions <- mrc %>%
  group_by(reg_name) %>%
  summarize(pop2016 = sum(pop2016))

The Power of Overlaying: Combining Layers for Deeper Insights

Overlaying is a technique that allows multiple data layers to be analyzed together, revealing relationships between different types of data. For example: - Overlaying economic data with geographic features: Understanding how natural barriers (e.g., rivers, mountains) impact economic clusters or transportation networks. - Combining social and environmental data: Correlating population density with climate data to study urban heat islands.

Overlaying reveals hidden correlations and offers a richer understanding of the spatial dimension of data.

Creating Custom Maps in QGIS

QGIS is designed for both simple and complex map creation. Here’s how to create a map: 1. Load Data Layers: Load the relevant vector and raster layers. 2. Print Layout: Use the Print Layout tool to arrange elements like maps, titles, and legends. 3. Export Map: Export your final map as an image, PDF, or SVG for presentation or publication.

In R, map creation is simplified with the ggplot2 package:

ggplot(data = mrc_proj) +
  geom_sf()

Managing and Transforming Coordinate Reference Systems (CRS)

CRS management ensures spatial alignment and accurate representation of data.

Transforming Coordinate Systems in QGIS

You can reproject layers to a new CRS using the Reproject Layer tool: 1. Select Layer: Choose the layer. 2. Reproject: Use the Reproject Layer tool to transform it to a new CRS. 3. Verify: Ensure the transformation was successful by checking the layer properties.

In R, reprojection can be done with sf:

mrc_proj <- st_transform(mrc, crs = 6622)
st_crs(mrc_proj)

Conclusion: Geospatial Data and Model Building

Geospatial data science transforms traditional models by adding the dimension of location, making them more representative of real-world phenomena. By integrating geospatial data into models, we can test and falsify theories more rigorously, moving closer to the truth. Models that ignore location overlook key variables, whereas those that incorporate spatial data provide a more holistic understanding of global patterns.

Working with Time Series Data in R

For those interested in working with time series data, here’s how you can reshape and visualize it in R:

year_labels <- c("SID74" = "1974 - 1978", "SID79" = "1979 - 1984")
nc_32119 %>%
  select(SID74, SID79) %>%
  pivot_longer(starts_with("SID")) -> nc_longer

ggplot() + 
  geom_sf(data = nc_longer, aes(fill = value), linewidth = 0.4) + 
  facet_wrap(~ name, ncol = 1, labeller = labeller(name = year_labels)) +
  scale_y_continuous(breaks = 34:36) +
  scale_fill_gradientn(colors = sf.colors(20)) +
  theme(panel.grid.major = element_line(color = "white"))

Exploring Raster and Vector Data in QGIS and R

Raster Data

Raster data is defined at each point of a grid covering the full extent of the data. It is often used for representing images, topographic maps, remote sensing data, and census data.

Vector Data

Vector data associates variables with discrete geometric objects located in space. It is commonly used to represent points, lines, polygons, and networks.

Combining Raster and Vector Data in QGIS and R

In both QGIS and R, raster and vector data can be combined in various ways. For example, you can query the raster at specific points or calculate aggregates over arbitrary regions.

In R, this can be done using the stars package:

library(stars)
st_extract(x, pts) # query at points
aggregate(x, st_buffer(pts, 500), FUN = mean) %>% st_as_sf() # aggregate over circles

2.4 Conclusion

This chapter introduced key tools in QGIS and R for geospatial data analysis, focusing on both vector and raster data. We explored basic data manipulation and visualization techniques, emphasizing reproducibility, transparency, and integration. The importance of understanding coordinate reference systems and projections was highlighted, alongside methods for transforming data between systems. The next chapter will delve into advanced spatial operations, spatial analysis, and the creation of interactive maps.

Data Repository

1. GeoPlatform

URL: geoplatform.gov
Provides a massive repository of geospatial data across various governmental levels (federal, state, county, local, tribal), with an emphasis on climate change, environmental, and policy datasets. Its broad scope and government backing make it one of the most valuable resources for spatial data.

2. EROS Center (U.S. Geological Survey)

URL: eros.usgs.gov
One of the most respected global repositories of aerial photography, satellite imagery, elevation data, and land cover datasets, the USGS EROS Center is a crucial institution for global geospatial research.

3. National Map Small-Scale Data Downloads

URL: nationalmap.gov/small_scale
Originally developed for the National Atlas project, this collection of datasets covers a broad range of topics, including transportation, land cover, boundaries, and water resources. These datasets provide excellent foundational information for U.S.-based projects.

4. TIGER Shapefile Downloads (U.S. Census Bureau)

URL: census.gov/geographies/mapping-files/time-series/geo/tiger-line-file.html
TIGER provides essential data for U.S. infrastructure and boundary-based projects, with shapefiles covering districts, roads, rivers, and more. The broad scope of national and state-level data makes it a must-have for projects based in the U.S.

5. Geospatial Data Gateway (U.S. Department of Agriculture)

URL: datagateway.nrcs.usda.gov
A useful source for county-based raster map files, especially orthoimagery. The U.S. Department of Agriculture’s gateway provides easily accessible geospatial datasets that are valuable for environmental and agricultural research.

6. DIVA-GIS

URL: diva-gis.org
Offering free geographic data for any country in the world, DIVA-GIS is widely used for environmental and biological projects. It provides a broad array of administrative and ecological data.

7. Global Administrative Areas (GADM)

URL: gadm.org
A significant resource for administrative boundary data worldwide, GADM supports multiple formats and coordinate systems, offering flexibility for diverse GIS applications.

8. Natural Earth

URL: naturalearthdata.com
Natural Earth provides global-scale vector and raster datasets for cartography and geospatial analysis. It is commonly used for creating visually appealing maps due to its focus on data simplicity and high quality.

9. MapCruzin

URL: mapcruzin.com
Known for providing free GIS maps and geospatial data, MapCruzin is a good resource for users interested in digital cartography, with a focus on U.S.-based data sets.

10. National Historical GIS

URL: nhgis.org
An essential resource for historical geospatial studies in the United States, providing census data and boundary files dating back to 1790.

11. GeoCommunicator (Bureau of Land Management)

URL: www.geocommunicator.gov
This platform focuses on the Public Land Survey System (PLSS) data and supports the mapping of federal land parcels. It’s particularly useful for land management and cadastral surveys in the U.S.

12. GeoCommunity – GIS Data Depot

URL: data.geocomm.com
GeoCommunity offers various types of raster data, including U.S. Geological Survey DRGs, DEMs, DOQQs, and FEMA Flood Data, making it a useful, though more specialized, resource.

13. GEOFABRIK

URL: download.geofabrik.de
GEOFABRIK provides OpenStreetMap shapefiles for streets and other geographic features worldwide. It is a highly specialized tool, particularly valuable for urban studies and infrastructure projects.

# Chapter 2: Making Sense of Spatial Data Using GIS & Geocoding and Georeferencing ## Introduction “Everything is related to everything else, but near things are more related than distant things.” – Waldo Tobler’s First Law of Geography. This simple yet profound law forms the foundation of geospatial analysis. When we model and analyze the world, location is not just a variable; it's a critical dimension that influences how things interact. Spatial proximity often drives relationships that might be missed in non-spatial models. For instance, predicting housing prices based purely on characteristics like size or age can miss the significant impact of location. A house next to a park or near public transportation will often command a higher price, even if it’s identical to one farther away. In this way, geospatial data allows us to enhance traditional models by incorporating the spatial dimension—a layer of truth that non-spatial models overlook. Geospatial data science sits at the crossroads of geography and data science, providing a framework for understanding the world through the lens of location. In today’s data-driven society, location matters more than ever. From smartphones equipped with GPS to vast networks of sensors, the explosion of geospatial data has revolutionized how we analyze global phenomena—whether it’s tracking climate change, analyzing economic trends, or mapping global migration patterns. This chapter will guide you through the foundational concepts and tools used in geospatial data science, focusing on how spatial data is represented, managed, and analyzed, primarily through the use of QGIS and R. It provides a comprehensive overview of essential tools available for geospatial data analysis, with a primary focus on QGIS and R. As geospatial data becomes increasingly prevalent across various fields—ranging from aerial photography and satellite imagery to census data and geolocated social media posts—understanding how to process and analyze this data is crucial. We will explore the two primary types of geospatial data: - **Raster data**: Represents variables at each point in a grid covering the data's full extent (e.g., satellite images). - **Vector data**: Associates variables with discrete geometric objects in space (e.g., city locations on a map). While QGIS offers a user-friendly graphical interface that simplifies many tasks, script-based processing in R remains essential for certain operations due to its reproducibility, transparency, and integration with other datasets. Throughout this chapter, we will demonstrate how to effectively use QGIS for various geospatial operations, with R code examples included for those interested in script-based analysis. ## The Prevalence of Geospatial Data Today, geospatial data is ubiquitous. From satellites capturing high-resolution images of the Earth to the growing availability of government data, organizations and researchers now have access to vast amounts of spatial data. The power of spatial data lies in its ability to show how things relate to one another geographically. For instance, analyzing the spatial distribution of economic activity alongside transportation networks can reveal how access to infrastructure impacts growth. ### Concepts This section offers an overview of the tools available in R for analyzing geolocated data. This type of data is becoming increasingly common in various domains, such as aerial photography, satellite imagery, census data, and geolocated social media posts. Geospatial data generally falls into two categories: - **Raster data**: Represents variables defined at each point of a grid covering the full extent of the data (e.g., satellite images). - **Vector data**: Associates variables (sometimes called attributes) with discrete geometric objects located in space (e.g., city locations on a map). ### Objectives of Script-Based Analysis While the graphical interface of GIS software like QGIS is intuitive, script-based processing in R offers several advantages: - **Reproducibility**: It is easy to repeat the analysis for new data by rerunning the script. - **Transparency**: Other researchers can reproduce the methods if they have access to the same programming language. - **Integration**: Spatial data can be extracted and merged with other datasets for statistical analysis within a single programming environment. ### Goals - Familiarize yourself with R packages for processing and visualizing vector and raster data (e.g., **sf** and **stars**). - Perform common data transformation operations using these packages. - Create more complex static maps (using **ggplot2**) and interactive maps (using **mapview**). ### Note on Packages - The set of packages available for spatial analysis in R has rapidly evolved. A few years ago, the **sp** and **raster** packages were the main tools for processing vector and raster data, respectively. **sf** and **stars** are part of a recent initiative to overhaul R's spatial tools ([https://www.r-spatial.org/](https://www.r-spatial.org/)). - The **sf** package represents spatial data frames with a standard format based on open-source geodatabases and integrates well with popular R packages for data manipulation and visualization (such as **dplyr** and **ggplot2**). - The **stars** package is also compatible with **ggplot2** and provides good support for raster "cubes" with non-spatial dimensions such as time. - The **raster** package and its successor **terra** (released in 2020) possess some features not present in **stars** and perform some operations more quickly. Therefore, it may be useful to learn them if your workflow involves complex operations on large rasters; see the documentation on [https://r-spatial.org/](https://r-spatial.org/) for more details. ## Working with Geospatial Data in QGIS and R Geospatial data science combines the principles of geography with data science to analyze phenomena through the dimension of location. In this chapter, we'll explore the foundational concepts and tools used in geospatial data science, with a focus on how spatial data is represented, analyzed, and visualized using QGIS and R. ### **Why Location Matters** **Waldo Tobler's First Law of Geography** states: “*Everything is related to everything else, but near things are more related than distant things.*” Location is a fundamental aspect of how the world functions, and by incorporating the **spatial dimension**, we can better understand relationships that may be overlooked in non-spatial models. This concept is central to geospatial analysis, which adds significant value to predictions, such as the relationship between housing prices and their proximity to infrastructure or amenities. ### **Types of Geospatial Data** Geospatial data comes in many forms, each suited for different types of analysis. Here are the most common data formats and types: #### **Data Formats**: | **Data Type** | **Non-Spatial Formats** | **Spatial Formats** | |----------------|-------------------------|------------------------| | **Text** | CSV, JSON, XML | GeoJSON, GML, KML | | **Binary** | PDF, XLS, ZIP | Shapefile, GeoPackage | | **Images** | TIFF, JPG, PNG | GeoTIFF, JPEG2000 | | **Databases** | SQLite, PostgreSQL | Spatialite, PostGIS | - **Shapefiles**: Widely used for storing geometries and attributes. - **GeoJSON**: Lightweight and web-friendly format for spatial data. - **GeoTIFF**: Common for raster data, such as satellite images. #### **Data Types**: - **Vector Data**: Points, lines, and polygons representing features like roads, cities, or administrative boundaries. - **Raster Data**: Gridded data that represents variables such as temperature or elevation, widely used in remote sensing. - **Tiles**: Subdivided raster or vector datasets, commonly used in web mapping. ### **Understanding Spatial Data: Geometry and Attributes** Spatial data is composed of two critical elements: 1. **Geometry**: This describes the physical shape and location of spatial features (points, lines, polygons). 2. **Attributes**: Additional data describing the features, such as population, area, or economic activity. ##### Example: GeoJSON Format A **GeoJSON** object represents spatial data by encoding geometry (coordinates) and attributes. Here is a simple GeoJSON for the city of Montreal: ```json { "type": "Feature", "geometry": { "type": "Point", "coordinates": [-73.5673, 45.5017] }, "properties": { "id": 1, "name": "Montreal" } } ``` This example includes both the geometry (longitude and latitude) and attributes (city name and ID). ### **Coordinate Reference Systems (CRS) and Map Projections** CRS are essential for representing the Earth’s 3D surface on 2D maps. Choosing the right CRS ensures accuracy in spatial analysis. #### **Map Projections**: - **Equal Earth projection**: Ideal for global maps, preserving area proportions. - **UTM (Universal Transverse Mercator)**: Commonly used for regional mapping due to minimized distortions in specific zones. ##### Example: Choosing a CRS For global maps, projections like **Equal Earth** offer visual appeal and minimal distortion. For local studies, country-specific CRSs, such as **EPSG:7755** for India, are used to maintain accuracy over small areas. ### **Tools for Geospatial Data Science: QGIS and R** #### **QGIS: A Practical Introduction** QGIS is an open-source GIS platform that offers a wide range of tools for **visualizing**, **manipulating**, and **analyzing** geospatial data. Its ability to handle multiple data formats, overlay layers, and visualize complex relationships makes it ideal for both beginners and advanced users. Key QGIS functions include: - **Data loading**: Use the **Data Source Manager** to load shapefiles, raster data, and other formats. - **Exploring attributes**: Tools like the **Identify** tool allow you to view and explore attribute data. - **Applying symbology**: Customize how data is visualized using the **Layer Styling Panel** to create meaningful maps. #### **R for Geospatial Analysis** R complements QGIS by offering advanced statistical capabilities. With packages like **sf** and **ggplot2**, R provides powerful tools for spatial analysis and data visualization. For example, the `sf` package allows seamless handling of spatial data, while `ggplot2` enables sophisticated data visualization. ##### Example: Loading and Exploring Data in R You can perform operations similar to QGIS using the `sf` package in R: ```r library(sf) mrc <- st_read("data/mrc.shp") st_bbox(mrc) st_crs(mrc) head(mrc) ``` ### **Working with Geospatial Data: Visualization and Manipulation** #### **Visualizing Geospatial Data in QGIS** In QGIS, you can visualize data by: 1. **Loading data**: Use the **Data Source Manager** to import data. 2. **Apply Symbology**: Open the **Layer Styling Panel** to customize the visualization based on attributes. In R, you can use `ggplot2` to create maps: ```r ggplot(data = mrc_proj) + geom_sf() ``` #### **Manipulating Geospatial Data in QGIS** QGIS offers various tools for data manipulation: - **Field Calculator**: Perform calculations or create new fields. - **Processing Toolbox**: Tools like buffering, clipping, and spatial joins. In R, similar data manipulation can be performed using `dplyr` and `sf`: ```r regions <- mrc %>% group_by(reg_name) %>% summarize(pop2016 = sum(pop2016)) ``` ### **The Power of Overlaying: Combining Layers for Deeper Insights** Overlaying is a technique that allows multiple data layers to be analyzed together, revealing relationships between different types of data. For example: - **Overlaying economic data with geographic features**: Understanding how natural barriers (e.g., rivers, mountains) impact economic clusters or transportation networks. - **Combining social and environmental data**: Correlating population density with climate data to study urban heat islands. Overlaying reveals hidden correlations and offers a richer understanding of the spatial dimension of data. ### **Creating Custom Maps in QGIS** QGIS is designed for both simple and complex map creation. Here’s how to create a map: 1. **Load Data Layers**: Load the relevant vector and raster layers. 2. **Print Layout**: Use the **Print Layout** tool to arrange elements like maps, titles, and legends. 3. **Export Map**: Export your final map as an image, PDF, or SVG for presentation or publication. In R, map creation is simplified with the `ggplot2` package: ```r ggplot(data = mrc_proj) + geom_sf() ``` ### **Managing and Transforming Coordinate Reference Systems (CRS)** CRS management ensures spatial alignment and accurate representation of data. #### **Transforming Coordinate Systems in QGIS** You can reproject layers to a new CRS using the **Reproject Layer** tool: 1. **Select Layer**: Choose the layer. 2. **Reproject**: Use the **Reproject Layer** tool to transform it to a new CRS. 3. **Verify**: Ensure the transformation was successful by checking the layer properties. In R, reprojection can be done with `sf`: ```r mrc_proj <- st_transform(mrc, crs = 6622) st_crs(mrc_proj) ``` ### Conclusion: Geospatial Data and Model Building Geospatial data science transforms traditional models by adding the dimension of location, making them more representative of real-world phenomena. By integrating geospatial data into models, we can **test and falsify** theories more rigorously, moving closer to the truth. Models that ignore location overlook key variables, whereas those that incorporate spatial data provide a more holistic understanding of global patterns. ### Working with Time Series Data in R For those interested in working with time series data, here’s how you can reshape and visualize it in R: ```r year_labels <- c("SID74" = "1974 - 1978", "SID79" = "1979 - 1984") nc_32119 %>% select(SID74, SID79) %>% pivot_longer(starts_with("SID")) -> nc_longer ggplot() + geom_sf(data = nc_longer, aes(fill = value), linewidth = 0.4) + facet_wrap(~ name, ncol = 1, labeller = labeller(name = year_labels)) + scale_y_continuous(breaks = 34:36) + scale_fill_gradientn(colors = sf.colors(20)) + theme(panel.grid.major = element_line(color = "white")) ``` ### Exploring Raster and Vector Data in QGIS and R #### Raster Data Raster data is defined at each point of a grid covering the full extent of the data. It is often used for representing images, topographic maps, remote sensing data, and census data. #### Vector Data Vector data associates variables with discrete geometric objects located in space. It is commonly used to represent points, lines, polygons, and networks. ### Combining Raster and Vector Data in QGIS and R In both QGIS and R, raster and vector data can be combined in various ways. For example, you can query the raster at specific points or calculate aggregates over arbitrary regions. In R, this can be done using the `stars` package: ```r library(stars) st_extract(x, pts) # query at points aggregate(x, st_buffer(pts, 500), FUN = mean) %>% st_as_sf() # aggregate over circles ``` ## Conclusion This chapter introduced key tools in QGIS and R for geospatial data analysis, focusing on both vector and raster data. We explored basic data manipulation and visualization techniques, emphasizing reproducibility, transparency, and integration. The importance of understanding coordinate reference systems and projections was highlighted, alongside methods for transforming data between systems. The next chapter will delve into advanced spatial operations, spatial analysis, and the creation of interactive maps. ## Data Repository {.unnumbered} ### 1. **GeoPlatform** **URL**: [geoplatform.gov](https://www.geoplatform.gov) Provides a massive repository of geospatial data across various governmental levels (federal, state, county, local, tribal), with an emphasis on climate change, environmental, and policy datasets. Its broad scope and government backing make it one of the most valuable resources for spatial data. ### 2. **EROS Center (U.S. Geological Survey)** **URL**: [eros.usgs.gov](https://eros.usgs.gov) One of the most respected global repositories of aerial photography, satellite imagery, elevation data, and land cover datasets, the USGS EROS Center is a crucial institution for global geospatial research. ### 3. **National Map Small-Scale Data Downloads** **URL**: [nationalmap.gov/small_scale](https://nationalmap.gov/small_scale) Originally developed for the National Atlas project, this collection of datasets covers a broad range of topics, including transportation, land cover, boundaries, and water resources. These datasets provide excellent foundational information for U.S.-based projects. ### 4. **TIGER Shapefile Downloads (U.S. Census Bureau)** **URL**: [census.gov/geographies/mapping-files/time-series/geo/tiger-line-file.html](https://www.census.gov/geographies/mapping-files/time-series/geo/tiger-line-file.html) TIGER provides essential data for U.S. infrastructure and boundary-based projects, with shapefiles covering districts, roads, rivers, and more. The broad scope of national and state-level data makes it a must-have for projects based in the U.S. ### 5. **Geospatial Data Gateway (U.S. Department of Agriculture)** **URL**: [datagateway.nrcs.usda.gov](https://datagateway.nrcs.usda.gov) A useful source for county-based raster map files, especially orthoimagery. The U.S. Department of Agriculture’s gateway provides easily accessible geospatial datasets that are valuable for environmental and agricultural research. ### 6. **DIVA-GIS** **URL**: [diva-gis.org](https://www.diva-gis.org) Offering free geographic data for any country in the world, DIVA-GIS is widely used for environmental and biological projects. It provides a broad array of administrative and ecological data. ### 7. **Global Administrative Areas (GADM)** **URL**: [gadm.org](https://www.gadm.org) A significant resource for administrative boundary data worldwide, GADM supports multiple formats and coordinate systems, offering flexibility for diverse GIS applications. ### 8. **Natural Earth** **URL**: [naturalearthdata.com](https://www.naturalearthdata.com) Natural Earth provides global-scale vector and raster datasets for cartography and geospatial analysis. It is commonly used for creating visually appealing maps due to its focus on data simplicity and high quality. ### 9. **MapCruzin** **URL**: [mapcruzin.com](https://www.mapcruzin.com) Known for providing free GIS maps and geospatial data, MapCruzin is a good resource for users interested in digital cartography, with a focus on U.S.-based data sets. ### 10. **National Historical GIS** **URL**: [nhgis.org](https://www.nhgis.org) An essential resource for historical geospatial studies in the United States, providing census data and boundary files dating back to 1790. ### 11. **GeoCommunicator (Bureau of Land Management)** **URL**: [www.geocommunicator.gov](https://www.geocommunicator.gov) This platform focuses on the Public Land Survey System (PLSS) data and supports the mapping of federal land parcels. It’s particularly useful for land management and cadastral surveys in the U.S. ### 12. **GeoCommunity – GIS Data Depot** **URL**: [data.geocomm.com](https://data.geocomm.com) GeoCommunity offers various types of raster data, including U.S. Geological Survey DRGs, DEMs, DOQQs, and FEMA Flood Data, making it a useful, though more specialized, resource. ### 13. **GEOFABRIK** **URL**: [download.geofabrik.de](https://download.geofabrik.de) GEOFABRIK provides OpenStreetMap shapefiles for streets and other geographic features worldwide. It is a highly specialized tool, particularly valuable for urban studies and infrastructure projects.