R for Conservation and Development Projects: A Primer for Practitioners

Cover
Half Title
Series Page
Title Page
Copyright Page
Dedication
Contents
Preface
1.
Introduction
	1.1.
What is R?
	1.2.
Why R?
	1.3.
Why this book?
	1.4.
What are development and conservation?
	1.5.
Science and decision making
	1.6.
Why data science is important
		1.6.1.
Monitoring and evaluation
		1.6.2.
Projects versus programmes
		1.6.3.
Project delivery versus research projects
	1.7.
The goal of this book
	1.8.
How this book is organised
	1.9.
How code is organised in this book
I:
Basics
	2.
Inference and Evidence
		2.1.
Inference
		2.2.
Study design
		2.3.
Evidence
		2.4.
What makes good data?
		2.5.
Recommended resources
		2.6.
Summary
	3.
Data integration in project management
		3.1.
Adaptive management cycles
		3.2.
The Deming cycle
			3.2.1.
Plan
				3.2.1.1.
Development of a project strategy and proposal
				3.2.1.2.
Proposal submission process
				3.2.1.3.
What is a logframe?
				3.2.1.4.
Logframe terminology
				3.2.1.5.
Pre-implementation planning
			3.2.2.
Train
			3.2.3.
Do
			3.2.4.
Check
			3.2.5.
Act
		3.3.
Challenges
		3.4.
Recommended resources
		3.5.
Summary
	4.
Getting started in R
		4.1.
Installing R
		4.2.
Installing RStudio
		4.3.
The R interface
			4.3.1.
The console
			4.3.2.
Version information
			4.3.3.
Writing code in the console
			4.3.4.
Script editors
			4.3.5.
Using the default script editor
			4.3.6.
Using RStudio
		4.4.
R as a calculator
		4.5.
How R works
			4.5.1.
Objects
			4.5.2.
Functions
				4.5.2.1.
Getting help on functions
			4.5.3.
Packages
				4.5.3.1.
Getting help on packages
		4.6.
Writing meaningful code
		4.7.
Reproducibility
		4.8.
Recommended resources
		4.9.
Summary
	5.
Introduction to data frames
		5.1.
Making data frames
		5.2.
Importing a data frame
		5.3.
Saving a data frame
		5.4.
Investigating a data frame
		5.5.
Other functions to examine an R object
		5.6.
Subsetting using the `[' and `]' operators
		5.7.
Descriptive statistics
		5.8.
Viewing data frames
		5.9.
Making a reproducible example
			5.9.1.
Reproducible example steps
		5.10.
Recommended resources
		5.11.
Summary
	6.
The Waihi project
		6.1.
The scenario
			6.1.1.
Why evidence is important
		6.2.
The data
			6.2.1. Description of condev data sets
		6.3.
Recommended resources
		6.4.
Summary
II:
First steps
	7.
ggplot2: graphing with the tidyverse
		7.1.
Why graph?
		7.2. The tidyverse package
		7.3.
The data
		7.4.
Graphing in R
			7.4.1.
Making a ggplot
			7.4.2.
Scatter plots
			7.4.3.
Bar plots
			7.4.4.
Histograms
			7.4.5.
Box plots
			7.4.6.
Polygons
			7.4.7.
Other common geoms
		7.5.
How to save a ggplot
		7.6.
Recommended resources
		7.7.
Summary
	8.
Customising a ggplot
		8.1.
Why customise a ggplot?
		8.2.
The packages
		8.3.
The data
		8.4.
Families of layers
		8.5.
Aesthetics properties
			8.5.1.
Settings aesthetics
			8.5.2.
A quick note about colour
			8.5.3.
Using aesthetics to distinguish groups
			8.5.4.
Using faceting to distinguish groups
		8.6.
Improving crowded graphs
		8.7.
Overlaying
		8.8.
Labels
		8.9.
Using the theme() function
			8.9.1.
The 4 elements
			8.9.2.
Rotating axis text
			8.9.3.
Spacing between axis and graph
			8.9.4.
In-built themes
		8.10.
Controlling axes
			8.10.1.
Tick marks
			8.10.2.
Axis limits
			8.10.3.
Forcing a common origin
			8.10.4.
Flipping axes
			8.10.5.
Forcing a plot to be square
			8.10.6.
Log scales and large numbers
		8.11.
Controlling legends
		8.12.
Recommended resources
		8.13.
Summary
	9.
Data wrangling
		9.1.
What is data wrangling?
		9.2.
The packages
		9.3.
The data
		9.4.
Pipes
		9.5.
Tibbles versus data frame
		9.6.
Subsetting
			9.6.1.
select()
			9.6.2.
lter()
		9.7.
Transforming
			9.7.1. group_by()
			9.7.2.
summarise()
			9.7.3.
mutate()
			9.7.4. adorn_totals()
		9.8.
Tidying
			9.8.1. pivot_wider()
			9.8.2. pivot_longer()
		9.9.
Ordering
			9.9.1. arrange()
			9.9.2. top_n()
		9.10.
Joining
		9.11.
Recommended resources
		9.12.
Summary
	10.
Data cleaning
		10.1.
Cleaning is more than correcting mistakes
		10.2.
The packages
		10.3.
The data
		10.4.
Changing names
			10.4.1. clean_names()
			10.4.2.
rename()
			10.4.3. fct_recode()
			10.4.4. str_replace all()
		10.5.
Fixing missing values
			10.5.1. fct_explicit na()
			10.5.2. replace_na()
			10.5.3.
replace()
			10.5.4. drop_na()
			10.5.5.
Cleaning a whole data set
		10.6.
Adding and dropping factor levels
			10.6.1. fct_drop()
			10.6.2. fct_expand()
			10.6.3.
Keeping empty levels in ggplot
		10.7.
Fusing duplicate columns
			10.7.1.
coalesce()
		10.8.
Organising factor levels
			10.8.1. fct_relevel()
			10.8.2. fct_reorder()
			10.8.3. fct_rev()
		10.9.
Anonymisation and pseudonymisation
			10.9.1. fct_anon()
		10.10.
Recommended resources
		10.11.
Summary
	11.
Working with dates and time
		11.1.
The two questions
		11.2.
The packages
		11.3.
The data
		11.4.
Formatting dates
			11.4.1. Formatting dates with lubridate
			11.4.2.
Formatting dates with base R
			11.4.3.
Numerical dates
		11.5.
Extracting dates
		11.6.
Time intervals
		11.7.
Time zones
			11.7.1.
The importance of time zones
			11.7.2.
Same times in di erent time zones
		11.8.
Replacing missing date components
		11.9.
Graphing: a worked example
			11.9.1.
Reordering a variable by a date
			11.9.2.
Summarising date-based data
			11.9.3. Date labels with scale_x _date()
		11.10.
Recommended resources
		11.11.
Summary
	12.
Working with spatial data
		12.1.
The importance of maps
		12.2.
The packages
		12.3.
The data
		12.4.
What is spatial data?
		12.5. Introduction to the sf package
			12.5.1. Reading data: st_read()
			12.5.2. Converting data: st_as_sf()
			12.5.3. Polygon area: st_area()
			12.5.4. Plotting maps: geom_sf()
			12.5.5. Extracting coordinates st_coordinates()
		12.6.
Plotting a world map with
			12.6.1. Filtering with flter()
		12.7.
Coordinate reference systems
			12.7.1. Finding the CRS of an object with st_crs()
			12.7.2. Transform the CRS with st_transform()
			12.7.3. Cropping with coord_sf()
		12.8.
Adding reference information
			12.8.1.
Adding a scale bar and north arrow
			12.8.2.
Positioning names with centroids
			12.8.3. Adding names with geom_text()
		12.9.
Making a chloropleth
		12.10.
Random sampling
		12.11. Saving with st_write()
		12.12. Rasters with the raster package
			12.12.1.
Loading rasters
			12.12.2.
Raster data
			12.12.3.
Plotting rasters
			12.12.4.
Basic raster calculations
			12.12.5.
Sampling
			12.12.6.
Extracting raster data from points
			12.12.7.
Turning data frames into rasters
			12.12.8.
Calculating distances
			12.12.9.
Masking
			12.12.10.
Cropping
			12.12.11.
Saving
			12.12.12.
Changing to a data frame
		12.13.
Recommended resources
		12.14.
Summary
	13.
Common R code mistakes and quirks
		13.1.
Making mistakes
		13.2.
The packages
		13.3.
The data
		13.4.
Capitalisation mistakes
		13.5.
Forgetting brackets
		13.6.
Forgetting quotation marks
		13.7.
Forgetting commas
		13.8.
Forgetting `+' in a ggplot
		13.9.
Forgetting to call a ggplot object
		13.10.
Piping but not making an object
		13.11.
Changing a factor to a number
		13.12.
Strings automatically read as factors
		13.13.
Summary
III:
Modelling
	14.
Basic statistical concepts
		14.1.
Variables and statistics
		14.2.
The packages
		14.3.
The data
		14.4.
Describing things which are variable
			14.4.1.
Central tendency
				14.4.1.1.
Mean
				14.4.1.2.
Median
			14.4.2.
Describing variability
				14.4.2.1.
Range
				14.4.2.2.
Standard deviation
				14.4.2.3.
Percentile range
			14.4.3.
Reporting central tendency and variability
			14.4.4.
Precision
		14.5.
Introducing probability
		14.6.
Probability distributions
			14.6.1.
Binomial distribution
				14.6.1.1.
Bernoulli distribution
			14.6.2.
Poisson distribution
			14.6.3.
Normal distribution
		14.7.
Random sampling
			14.7.1.
Simple random sampling
			14.7.2.
Strati ed random sampling
		14.8.
Modelling approaches
			14.8.1.
Null hypothesis testing
			14.8.2.
Information-theoretics
			14.8.3.
Bayesian approaches
			14.8.4.
Machine learning
		14.9.
Undertting and overtting
		14.10.
Recommended resources
		14.11.
Summary
	15.
Understanding linear models
		15.1.
Regression versus classi cation
		15.2.
The packages
		15.3.
The data
		15.4. Graphing a y variable
		15.5.
What is a linear model?
			15.5.1.
How to draw a linear model from an equation
		15.6.
Predicting the response variable
		15.7.
Formulating hypotheses
		15.8.
Goodness-of-fit
			15.8.1.
Residuals
			15.8.2.
Correlation
		15.9.
Making a linear model in R
		15.10.
Introduction to model selection
			15.10.1. Estimating the number of parameters: K
			15.10.2. Goodness of t: L
		15.11.
Doing model selection in R
			15.11.1.
Interpreting an AIC Table
				15.11.1.1.
Evidence ratios
				15.11.1.2.
Keep in mind
		15.12.
Understanding coe cients
		15.13.
Model equations and prediction
			15.13.1.
Dummy variables and a design matrix
			15.13.2. Plotting a prediction with geom_abline()
			15.13.3.
Automatic prediction
		15.14.
Understanding a model summary
		15.15.
Standard errors and con dence intervals
			15.15.1. Confidence intervals for model predictions
		15.16.
Model diagnostics
			15.16.1.
Still problems?
		15.17.
Log transformations
			15.17.1.
What are logarithms?
			15.17.2.
Logarithms and zero
		15.18.
Simulation
			15.18.1.
Making a for() loop
			15.18.2.
Example simulation
		15.19.
Reporting modelling results
		15.20.
Summary
	16.
Extensions to linear models
		16.1.
Building upon linear models
		16.2.
The packages
		16.3.
The data
		16.4.
Multiple regression
			16.4.1.
Additive versus interaction models
			16.4.2.
Visualising multiple regression
				16.4.2.1.
Visualising continuous variables
				16.4.2.2.
A visualisation trick
			16.4.3.
Colinearity and multicolinearity
		16.5.
Most statistical tests are linear models
			16.5.0.1.
One sample t-test
			16.5.0.2.
Independent t-test
			16.5.0.3.
One-way ANOVA
			16.5.0.4.
Two-way ANOVA
			16.5.0.5.
ANCOVA
		16.6.
Generalised linear models
			16.6.0.1.
The importance of link functions
			16.6.1.
Gaussian distribution (normal distribution)
			16.6.2.
Binomial distribution for logistic regression
				16.6.2.1. Confidence intervals with link functions
			16.6.3.
Poisson regresssion for count data
				16.6.3.1.
Diagnostics for GLM
				16.6.3.2.
Under and over-dispersion
			16.6.4.
Chi-squared tests
				16.6.4.1.
Data preparation
				16.6.4.2.
Chi-squared goodness-oft test
				16.6.4.3.
Test of independence
		16.7.
Other related modelling approaches
			16.7.1.
Repeated measures
				16.7.1.1.
Pairwise t-test
			16.7.2.
Cumulative link models
			16.7.3.
Beta regression
			16.7.4.
Non-parametric approaches
			16.7.5.
Advantages of non-parametric tests
		16.8.
Recommended resources
		16.9.
Summary
	17.
Introduction to clustering and classification
		17.1.
Clustering and Classi cation
		17.2.
The packages
		17.3.
The data
		17.4.
Supervised versus unsupervised learning
			17.4.1.
Why learn classi cation and clustering
		17.5.
Clustering
			17.5.1.
Hierarchical clustering
			17.5.2.
Dimension reduction
		17.6.
Classification
			17.6.1.
Classification trees
			17.6.2. How classification trees work
			17.6.3.
Model evaluation
				17.6.3.1.
Pruning a tree
			17.6.4.
k-fold cross-validation
			17.6.5.
Accuracy versus interpretability
		17.7.
Recommended resources
		17.8.
Summary
	18.
Reporting and worked examples
		18.1.
Writing the project report
		18.2.
The packages
		18.3.
The data
		18.4.
Reporting training data
		18.5.
Prosecution results
		18.6.
Fish pond study
	19.
Epilogue
A.
Appendix: step-wise statistical calculations
	A.1.
How to approach an equation
	A.2.
Data
	A.3.
The standard deviation
	A.4.
Model coe cients
		A.4.1.
Slope
		A.4.2.
Intercept
		A.4.3.
r2 (Pearson's correlation coeffcient)
		A.4.4.
AIC and AICc
Index