OLD_3.Linear_regression.R

# Hannah-Marie Lamle
# LSAT habitat suitability modelling
# Simple linear regressions first
# Creation date of this document: 11/6/2025

# This document was created during cleanup of this project repository to better
# aid in data analyzation for the future. 

library(tidyverse)
library(ggplot2)
library(readxl)
library(dplyr)
library(patchwork)
library(ggpubr)
library(grid)


# -------------------- Old code: not using ---------------------------------


#ggplot(rawdata25, aes(Turf_length, Sediment_depth))+      # modeling sediment depth (y) as 
#  geom_point()+                                           # function of turf length (x)
#  geom_smooth(method = "lm", formula = y ~ x)
#model <- lm(Sediment_depth ~ Turf_length, data = Comparison_metrics_South_Canyon)
#summary(model)

#ggplot(Comparison_metrics_South_Canyon, aes(Turf_length, avg_slope))+           # modelling slope (y) as a function
#  geom_point()+                                                                 # of turf legnth (x)
#  geom_smooth(method = "lm", formula = y ~ x)
#model <- lm(avg_slope ~ Turf_length, data = Comparison_metrics_South_Canyon)
#summary(model)

#ggplot(Comparison_metrics_South_Canyon, aes(Turf_length, avg_rugosity))+        # modelling rugosity (y) as function
#  geom_point()+                                                                 # of turf length (x)
#  geom_smooth(method = "lm", formula = y ~ x)
#model <- lm(avg_rugosity ~ Turf_length, data = Comparison_metrics_South_Canyon)
#summary(model)

#ggplot(Comparison_metrics_South_Canyon, aes(avg_slope, avg_rugosity))+          # modelling rugosity (y) as function 
#  geom_point()+                                                                 # of turf length (x)
#  geom_smooth(method = "lm", formula = y ~ x)
#model <- lm(avg_rugosity ~ avg_slope, data = Comparison_metrics_South_Canyon)
#summary(model)

# From Rolo: we are trying to see if microstructure taken from Viscore/GIS can 
# reliably predict LSAT metrics (sediment depth, turf length). 

# Therefore, need to flip the x and y axes and model LSAT as a function of complexity

# Feedback from Rolo incorporated, here is how microstructure ranges are altering 
# turf depth and sediment depth: 


# --------------------------- Linear Regressions: 25 ---------------------------


# regular rugosity vs turf length
rugo25_turf <- ggplot(rawdata25, aes(x = avg_rugo25, y = `turf_ length_mm`))+   # Modelling turf length (y) as function of rugosity (x)
  geom_point()+
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("25cm Box")
model <- lm(`turf_ length_mm` ~ avg_rugo25, data = rawdata25)
summary(model)
rugo25_turf


# regular rugosity vs sediment depth
rugo25_sediment <- ggplot(rawdata25, aes(x = avg_rugo25,  y = sed_depth_mm))+     # modelling sediment depth (y) as function
  geom_point()+                                                                 # of rugosity (x)
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")
model <- lm(sed_depth_mm ~ avg_rugo25, data = rawdata25)
summary(model)
rugo25_sediment

# 0-1 rugosity vs turf length
#ggplot(rawdata25, aes(rugo25_A, y = `Turf length (mm)`))+     # modelling sediment depth (y) as function
#  geom_point()+                                                                 # of rugosity (x)
#  geom_smooth(method = "lm", formula = y ~ x)+
#  xlab("Rugosity on 0-1 scale")
#model <- lm(`Turf length (mm)` ~ rugo25_A, data = rawdata25)
#summary(model)

# 0-1 rugosity vs sediment depth
#ggplot(rawdata25, aes(rugo25_A, y = `Sediment depth (mm)`))+     # modelling sediment depth (y) as function
#  geom_point()+                                                                 # of rugosity (x)
#  geom_smooth(method = "lm", formula = y ~ x)+
#  xlab("Rugosity on 0-1 scale")
#model <- lm(`Sediment depth (mm)` ~ rugo25_A, data = rawdata25)
#summary(model)


# slope vs turf length
slope25_turf <- ggplot(rawdata25, aes(slope, y = `turf_ length_mm`))+           # modelling turf length (y) as function 
  geom_point()+                                                                 # of slope
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")+
  ggtitle("25cm Box")
model <- lm(`turf_ length_mm` ~ MEAN, data = rawdata25)
summary(model)
slope25_turf

# slope vs sediment depth 
slope25_sediment <- ggplot(rawdata25, aes(slope, y = sed_depth_mm))+        # modelling sediment depth (y) as function
  geom_point()+                                                                 # of slope (x)
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")
model <- lm(sed_depth_mm ~ MEAN, data = rawdata25)
summary(model)
slope25_sediment


# --------------------------- Linear Regressions: 50 ---------------------------


# regular rugosity vs turf length
rugo50_turf <- ggplot(rawdata50, aes(x = avg_rugo50, y = `turf_ length_mm`))+   # Modelling turf length (y) as function of rugosity (x)
  geom_point()+
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("50cm Box")
model <- lm(`turf_ length_mm` ~ avg_rugo50, data = rawdata50)
summary(model)


# regular rugosity vs sediment depth
rugo50_sediment <- ggplot(rawdata50, aes(x = avg_rugo50,  y = sed_depth_mm))+     # modelling sediment depth (y) as function
  geom_point()+                                                                 # of rugosity (x)
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")
model <- lm(sed_depth_mm ~ avg_rugo50, data = rawdata50)
summary(model)


# 0-1 rugosity vs turf length
#ggplot(rawdata50, aes(rugo50_A, y = `Turf length (mm)`))+     # modelling sediment depth (y) as function
#  geom_point()+                                                                 # of rugosity (x)
#  geom_smooth(method = "lm", formula = y ~ x)+
#  xlab("Rugosity on 0-1 scale")
#model <- lm(`Turf length (mm)` ~ rugo50_A, data = rawdata50)
#summary(model)

# 0-1 rugosity vs sediment depth
#ggplot(rawdata50, aes(rugo50_A, y = `Sediment depth (mm)`))+     # modelling sediment depth (y) as function
#  geom_point()+                                                                 # of rugosity (x)
#  geom_smooth(method = "lm", formula = y ~ x)+
#  xlab("Rugosity on 0-1 scale")
#model <- lm(`Sediment depth (mm)` ~ rugo50_A, data = rawdata50)
#summary(model)


# slope vs turf length
slope50_turf <- ggplot(rawdata50, aes(slope, y = `turf_ length_mm`))+           # modelling turf length (y) as function 
  geom_point()+                                                                 # of slope
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")+
  ggtitle("50cm Box")
model <- lm(`turf_ length_mm` ~ MEAN, data = rawdata50)
summary(model)


# slope vs sediment depth 
slope50_sediment <- ggplot(rawdata50, aes(slope, y = sed_depth_mm))+        # modelling sediment depth (y) as function
  geom_point()+                                                                 # of slope (x)
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")
model <- lm(sed_depth_mm ~ MEAN, data = rawdata50)
summary(model)


# --------------------------- Linear Regressions: 100 ---------------------------


# regular rugosity vs turf length
rugo100_turf <- ggplot(rawdata100, aes(x = avg_rugo100, y = `turf_ length_mm`))+   # Modelling turf length (y) as function of rugosity (x)
  geom_point()+
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("100cm Box")
model <- lm(`turf_ length_mm` ~ avg_rugo100, data = rawdata100)
summary(model)


# regular rugosity vs sediment depth
rugo100_sediment <- ggplot(rawdata100, aes(x = avg_rugo100,  y = sed_depth_mm))+     # modelling sediment depth (y) as function
  geom_point()+                                                                 # of rugosity (x)
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")
model <- lm(sed_depth_mm ~ avg_rugo100, data = rawdata100)
summary(model)


# 0-1 rugosity vs turf length
#ggplot(rawdata100, aes(rugo100_A, y = `Turf length (mm)`))+     # modelling sediment depth (y) as function
#  geom_point()+                                                                 # of rugosity (x)
#  geom_smooth(method = "lm", formula = y ~ x)+
#  xlab("Rugosity on 0-1 scale")
#model <- lm(`Turf length (mm)` ~ rugo100_A, data = rawdata100)
#summary(model)

# 0-1 rugosity vs sediment depth
#ggplot(rawdata100, aes(rugo100_A, y = `Sediment depth (mm)`))+     # modelling sediment depth (y) as function
#  geom_point()+                                                                 # of rugosity (x)
#  geom_smooth(method = "lm", formula = y ~ x)+
#  xlab("Rugosity on 0-1 scale")
#model <- lm(`Sediment depth (mm)` ~ rugo100_A, data = rawdata100)
#summary(model)


# slope vs turf length
slope100_turf <- ggplot(rawdata100, aes(slope, y = `turf_ length_mm`))+           # modelling turf length (y) as function 
  geom_point()+                                                                 # of slope
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")+
  ggtitle("100cm Box")
model <- lm(`turf_ length_mm` ~ MEAN, data = rawdata100)
summary(model)


# slope vs sediment depth 
slope100_sediment <- ggplot(rawdata100, aes(slope, y = sed_depth_mm))+        # modelling sediment depth (y) as function
  geom_point()+                                                                 # of slope (x)
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")
model <- lm(sed_depth_mm ~ MEAN, data = rawdata100)
summary(model)


# ---------------------------- Plotting figs ------------------------

## ------------ Rugosity metrics -----------------

p1 <- ggplot(rawdata25, aes(x = avg_rugo25, y = `Turf length (mm)`))+   
  geom_point()+
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("25cm microquadrat")+
  theme_classic() <- ggplot(rawdata25, aes(x = avg_rugo25, y = `Turf length (mm)`))+   
  geom_point()+
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("25cm microquadrat")+
  theme_classic()

p2 <- ggplot(rawdata25, aes(x = avg_rugo25,  y = `Sediment depth (mm)`))+     
  geom_point()+                                                               
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("25cm microquadrat")+
  theme_classic()

p3 <- ggplot(rawdata50, aes(x = avg_rugo50, y = `Turf length (mm)`))+   # Modelling turf length (y) as function of rugosity (x)
  geom_point()+
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("50cm microquadrat")+
  theme_classic()

p4 <- ggplot(rawdata50, aes(x = avg_rugo50,  y = `Sediment depth (mm)`))+     
  geom_point()+                                                                
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("50cm microquadrat")+
  theme_classic()

p5 <- ggplot(rawdata100, aes(x = avg_rugo100, y = `Turf length (mm)`))+   # Modelling turf length (y) as function of rugosity (x)
  geom_point()+
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("100cm quadrat")+
  theme_classic()

p6 <- ggplot(rawdata100, aes(x = avg_rugo100,  y = `Sediment depth (mm)`))+   
  geom_point()+                                                                 
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Rugosity")+
  ggtitle("100cm Quadrat")+
  theme_classic()

(p1 | p2) / (p3 | p4) / (p5 | p6) 

# use this instead, it just uses the graphs created above rather than rewriting them: 

(rugo25_turf | rugo25_sediment) / (rugo50_turf | rugo50_sediment) / (rugo100_turf | rugo100_sediment) +
  plot_annotation(
    title = "Rugosity Results",
  )



## ------------ Slope Metrics -----------------

p1 <- ggplot(rawdata25, aes(MEAN, y = `Turf length (mm)`))+          
  geom_point()+                                                                 
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")+
  ggtitle("25cm microquadrat")+
  theme_classic()

p2 <- ggplot(rawdata25, aes(MEAN, y = `Sediment depth (mm)`))+        
  geom_point()+                                                                
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")+
  ggtitle("25cm Quadrat")+
  theme_classic()

p3 <- ggplot(rawdata50, aes(MEAN, y = `Turf length (mm)`))+      
  geom_point()+                                                                
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")+
  ggtitle("50cm microquadrat")+
  theme_classic()

p4 <- ggplot(rawdata50, aes(MEAN, y = `Sediment depth (mm)`))+     
  geom_point()+                                                                 
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")+
  ggtitle("50cm Quadrat")+
  theme_classic()

p5 <- ggplot(rawdata100, aes(MEAN, y = `Turf length (mm)`))+       
  geom_point()+                                                                 
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")+
  ggtitle("100cm Quadrat")+
  theme_classic()

p6 <- ggplot(rawdata100, aes(MEAN, y = `Sediment depth (mm)`))+       
  geom_point()+                                                                 
  geom_smooth(method = "lm", formula = y ~ x)+
  xlab("Average Slope (degrees)")+
  theme_classic()

(p1 | p2) / (p3 | p4) / (p5 | p6) 

# use this instead, it just uses the graphs created above rather than rewriting them: 

(slope25_turf | slope25_sediment) / (slope50_turf | slope50_sediment) / (slope100_turf | slope100_sediment) +
  plot_annotation(
    title = "Slope Results",
  )