# load libraries
library(ggplot2)
library(dplyr)
library(tidyr)

# ggplot format
legend_theme <- theme(plot.title = element_text(size=20, hjust = 0.5, face="bold"), 
                      axis.title.x = element_text(size=14, face="bold"),
                      axis.title.y = element_text(size=14, face="bold"),
                      legend.title=element_text(size=14),
                      legend.text=element_text(size=12))

theme <- theme(plot.title = element_text(size=20, hjust = 0.5, face="bold"),
               axis.title.x = element_text(size=14, face="bold"),
               axis.title.y = element_text(size=14, face="bold"),
               axis.text.x=element_blank(),
               axis.text.y=element_blank())

# read data
data <- read.csv("acceleration.csv")

# print(data)
head(data, 10)

# markers <- subset(data, Key == 61)
# markers <- subset(data, Key %in% c(61)) # with c(x,y,z,...) a list of matching keys
markers <- c(t(subset(data, Key == 61, select=Time)))
# markers <- subset(data, Key == 61, select=c(Time,Value)) # keep Time and Value column

move <- data
move <- subset(data, Time > markers[11] & Time < markers[12])

gg <- ggplot(move)
gg <- gg + geom_line(aes(x=Time, y=Value, color=factor(Key, labels = c("X", "Y", "Z"))))  # with factor color is taken discrete
gg <- gg + labs(title = "Acceleration", x = "Time (s)", y = "Acceleration (m/s²)", color="Direction")
gg <- gg + legend_theme
print(gg)

# library(plotly)
# ggplotly(gg)

#ggplot(subset(move, Key %in% c("11", "12"))) +
#  geom_line(aes(x=Time, y=Value, color=factor(Key))) + # factor makes color discrete
#  legend_theme

# extract accelerations
xdata <- subset(move, Key == 11, select=c("Time","Value"))
ydata <- subset(move, Key == 12, select=c("Time","Value"))
zdata <- subset(move, Key == 13, select=c("Time","Value"))

# substract mean (eliminate acceleration due to gravity)
xdata$Acceleration = xdata$Value - mean(xdata$Value)
ydata$Acceleration = ydata$Value - mean(ydata$Value)
zdata$Acceleration = zdata$Value - mean(zdata$Value)

# integrate twice to optain position data
xdata$Position = c(0, cumsum(cumsum(xdata$Acceleration[-1]*diff(xdata$Time))*diff(xdata$Time)))
ydata$Position = c(0, cumsum(cumsum(ydata$Acceleration[-1]*diff(ydata$Time))*diff(ydata$Time)))
zdata$Position = c(0, cumsum(cumsum(zdata$Acceleration[-1]*diff(zdata$Time))*diff(zdata$Time)))

# Visualize results
ggplot() + 
  geom_point(aes(x=xdata$Time,y=xdata$Position)) + 
  labs(title = "X-Movement", x = "Time", y = "X") + theme

ggplot() + 
  geom_point(aes(x=ydata$Time,y=ydata$Position)) + 
  labs(title = "Y-Movement", x = "Time", y = "Y") + theme

ggplot() + 
  geom_point(aes(x=zdata$Time,y=zdata$Position)) + 
  labs(title = "Z-Movement", z = "Time", y = "Z") + theme

ggplot() + 
  geom_point(aes(x=xdata$Position,y=ydata$Position)) + 
  labs(title = "X-Y-Movement", x = "X", y = "Y") + theme

ggplot() + 
  geom_point(aes(x=ydata$Position,y=zdata$Position)) + 
  labs(title = "X-Z-Movement", x = "X", y = "Z") + theme

ggplot() + 
  geom_point(aes(x=xdata$Position ,y=zdata$Position)) + 
  labs(title = "Y-Z-Movement", x = "Y", y = "Z") + theme

library(rgl)
plot3d(xdata$Position,ydata$Position,zdata$Position, box = TRUE, axes = TRUE)

library(plotly)
df <- data.frame(x = xdata$Position, y = ydata$Position, z = zdata$Position)
plot_ly(df, x = ~x, y = ~y, z = ~z)