Computer logs (orig).Rmd

---
title: "Computer Experiment"
author: "Ryan"
date: "03/23/2023"
output: word_document
---
########################### HWinfo64 ##############################################

########## Gaming ##########


```{r}
library(tidyverse)
library(readxl)
library(lubridate) #for date extraction and manipulation
library(ggthemes)

hwinfo_ow <- read_excel(path = "Data/hwinfo_test.xlsx")

#the -2 removes the last 2 observations in the data set which are just descriptors of the variables 
headed <- head(hwinfo_ow, -2)

#Converting all the variables in the dataset, expect for date, into numeric variables
headed_num <- as.data.frame(lapply(headed[,-1], as.numeric))

#imputing date variable back into the headed_num dataframe
headed_num <- cbind(headed$`Date/Time`, headed_num)

#Rename the date column we just added back above 
names(headed_num)[names(headed_num) == "headed$`Date/Time`"] <- "DateTime"

# Convert DateTime variable to POSIXct format
headed_num$DateTime <- ymd_hms(headed_num$DateTime)

# Convert DateTime variable to numeric representing seconds since initialization point
headed_num$seconds <- as.numeric(headed_num$DateTime - min(headed_num$DateTime))

#removing unimportant datasets
rm(hwinfo_ow, headed)
```


```{r}
#Getting the basic summary statistics for all of the relevant variables in the dataframe 
head_sum <- headed_num %>% select(-DateTime, -seconds)

#Applying the summary function to every variable in our modified head_sum dataframe
summary <- lapply(head_sum, summary)

#Getting the variable names from the summary list so that I can index into specific variable of interst 
head_sum_names_list <- names(summary)

summary[["GPU.Temperature...C."]]

summary
```


#####this code works great but the varaibles need to be group by their units otherwise the boxplot looks horrible 

```{r}
#Trying to visualize the statistics summary above into boxplots, 

data_long <- pivot_longer(head_sum, cols = head_sum_names_list, names_to = "variable")

# Create a boxplot using ggplot2
ggplot(data_long, aes(x = variable, y = value)) +
  geom_boxplot() +
  labs(title = "Boxplot of Multiple Groups",
       x = "Group", y = "Value")

```



```{r}
#conducting a t-test (difference in means) between the 2 variables i'm using to measure Cpu temps

# Conduct the t-test CPU TEMP
t_test_cpu <- t.test(headed_num$CPU.Die..average....C., headed_num$CPU..Tctl.Tdie....C.)
t_test_cpu

#t test beteen cpu and gpu power draw
t_test_gpu <- t.test(headed_num$CPU.Package.Power..W., headed_num$GPU.PPT..W.)
t_test_gpu

#The first t-test is comparing the means of two variables: headed_num$CPU.Die..average....C. and headed_num$CPU..Tctl.Tdie....C.. The results indicate that there is a statistically significant difference between the means of these two variables (t = -4.4571, p < 0.001). The negative t-value indicates that the mean of headed_num$CPU.Die..average....C. is smaller than the mean of headed_num$CPU..Tctl.Tdie....C.. The 95% confidence interval (-3.568860, -1.369849) suggests that we can be 95% confident that the true difference between the means of these two variables falls within this range.

#The second t-test is comparing the means of two different variables: headed_num$CPU.Package.Power..W. and headed_num$GPU.PPT..W.. The results indicate that there is a highly statistically significant difference between the means of these two variables (t = -33.836, p < 0.001). The negative t-value indicates that the mean of headed_num$CPU.Package.Power..W. is smaller than the mean of headed_num$GPU.PPT..W.. The 95% confidence interval (-59.55398, -52.92840) suggests that we can be 95% confident that the true difference between the means of these two variables falls within this range.
```


```{r}
headed_num %>% ggplot(aes(seconds, GPU.Temperature...C.)) +
  geom_line(aes(y = GPU.Temperature...C., color = "GPU.Temperature...C.")) +
  #geom_smooth(aes(y = CPU.Die..average....C., color = "CPU.Die"), se = FALSE, linetype = "dashed") +
  
  geom_line(aes(y = CPU..Tctl.Tdie....C., color = "CPU Tctl Tdie")) +
  #geom_smooth(aes(y = CPU..Tctl.Tdie....C., color = "CPU.Tctl.Tdie"), se = FALSE, linetype = "dashed") +
  
  labs(
    y = "Temperature (*C)",
    x = "Time - Minute:Seconds", 
    title = "Comparing different measures of CPU Temp over time"
  ) +
  
  scale_x_continuous(
    breaks = seq(0, max(headed_num$seconds), by = 20),
    
    #sprintf() is a function in R that allows you to format strings with placeholders for variables. It takes two arguments: the format string and the variables to substitute into the string. The format string is a character string that includes placeholders for the variables.
    
    labels = function(x) sprintf("%02d:%02d", floor(x/60), x %% 60)
  ) +
  
  #scale_y_continuous(breaks = seq(40, 80, by = 5),
    #minor_breaks = seq(40, 80, by = 2.5),
    #limits = c(40, 80)) +
  
  scale_color_manual(name = "CPU Temp Measures",
                     values = c("GPU.Temperature...C." = "red", "CPU Tctl Tdie" = "blue")) +
  
  theme_bw() +
  
  theme(panel.grid.minor.y = element_line(color = "gray", linetype = "dotted"),
       legend.position = c(0.85, 0.88),
       legend.title = element_text(size = 10.5), 
       legend.text = element_text(size = 10),
       legend.background = element_rect(fill = "grey99", color = "black", size = 0.5)) #makes a box around the legend and adds a fill! 
```


```{r}
headed_num %>% ggplot(aes(seconds)) +
  geom_line(aes(y = CPU.Package.Power..W., color = "CPU")) +
  
  geom_line(aes(y = GPU.PPT..W., color = "GPU")) +
  
  
  labs(
    y = "Power Draw in Watts",
    x = "Time - (Minute:Seconds)", 
    title = "Comparing Power Consumption of CPU and GPU Over Time"
  ) +
  
  scale_y_continuous(breaks = seq(0, 300, by = 50),
                     minor_breaks = seq(0, 300, by = 25),
                     limits = c(0, 300)) +
  
  scale_x_continuous(
    breaks = seq(0, max(headed_num$seconds), by = 10),
    
    #sprintf() is a function in R that allows you to format strings with placeholders for variables. It takes two arguments: the format string and the variables to substitute into the string. The format string is a character string that includes placeholders for the variables.
    
    labels = function(x) sprintf("%2d:%02d", floor(x/60), x %% 60)
  ) +
  
  scale_color_manual(name = "Component", values = c("CPU" = "red", "GPU" = "blue")) +
  theme_clean() +
  
  theme(
       panel.grid.minor.y = element_line(color = "gray", linetype = "dotted"),
       legend.position = c(0.80, 0.92),
       legend.direction = "horizontal",
       legend.title = element_text(size = 10.5), 
       legend.text = element_text(size = 10),
       legend.background = element_rect(fill = "grey99", color = "black", size = 0.5)) #makes a box around the legend and adds a fill! 
```


########## Synthetic ##########

###### CPU ######
```{r}
library(tidyverse)
library(readxl)
library(lubridate) #for date extraction and manipulation
library(ggthemes)

hwinfo_cinebench <- read_excel("C:/Users/Ryan/Coding Projects/Computer Experiment/Test Com Data.xlsx", sheet = "hwinfo_cinebench")

#the -2 removes the last 2 observations in the data set which are just descriptors of the variables 
headed <- head(hwinfo_cinebench, -2)

#Converting all the variables in the dataset, expect for date, into numeric variables
headed_num <- as.data.frame(lapply(headed[,-1], as.numeric))

#imputing date variable back into the headed_num dataframe
headed_num <- cbind(headed$`Date/Time`, headed_num)

#Rename the date column we just added back above 
names(headed_num)[names(headed_num) == "headed$`Date/Time`"] <- "DateTime"

# Convert DateTime variable to POSIXct format
headed_num$DateTime <- ymd_hms(headed_num$DateTime)

# Convert DateTime variable to numeric representing seconds since initialization point
headed_num$seconds <- as.numeric(headed_num$DateTime - min(headed_num$DateTime))

#removing unimportant datasets
rm(hwinfo_cinebench, headed)
```


```{r}
#Getting the basic summary statistics for all of the relevant variables in the dataframe 
head_sum <- headed_num %>% select(-DateTime, -seconds)

#Applying the summary function to every variable in our modified head_sum dataframe
summary <- lapply(head_sum, summary)

#Getting the variable names from the summary list so that I can index into specific variable of interst 
head_sum_names_list <- names(summary)

summary[["GPU.Temperature...C."]]

summary
```


```{r}
#Temperature and power consumption graph 

headed_num %>% ggplot(aes(seconds)) +
  geom_line(aes(y = CPU.Package.Power..W., color = "CPU.Package.Power..W.")) +
  geom_line(aes(y = CPU..Tctl.Tdie....C., color = "CPU Tctl Tdie")) +
  
  geom_text(aes(label = "Cinebench Score: 19406"), x = 54, y = 50) +
  
  labs(
    x = "Time - Minute:Seconds", 
    title = "Comparing different measures of CPU Temp over time"
  ) +
  
  scale_x_continuous(
    breaks = seq(0, max(headed_num$seconds), by = 20),
    
    #sprintf() is a function in R that allows you to format strings with placeholders for variables. It takes two arguments: the format string and the variables to substitute into the string. The format string is a character string that includes placeholders for the variables.
    
    labels = function(x) sprintf("%02d:%02d", floor(x/60), x %% 60)
  ) +
  
  scale_y_continuous(name = "Watts" ,
    sec.axis = sec_axis(~., name =  "Temperature (*C)")) +
    #breaks = seq(40, 80, by = 5),
    #minor_breaks = seq(40, 80, by = 2.5),
    #limits = c(40, 80)) +
  
  scale_color_manual(name = "CPU Temp Measures",
                     values = c("CPU Tctl Tdie" = "red", "CPU.Package.Power..W." = "blue")) +
  
  theme_bw() +
  
  theme(panel.grid.minor.y = element_line(color = "gray", linetype = "dotted"),
       legend.position = c(0.65, 0.40),
       legend.title = element_text(size = 10.5), 
       legend.text = element_text(size = 10),
       legend.background = element_rect(fill = "grey99", color = "black", size = 0.5)) + #makes a box around the legend and adds a fill! 
  
  guides(fill = guide_legend(reverse = TRUE)) 

#The "cairo = FALSE" parameter should get rid of the anti aliasing on the cinebench text
#ggsave(filename = "my_plot.png", dpi = 300, cairo = FALSE)
```


```{r}
#Clock Speed graph
names(headed_num)

headed_num %>% ggplot(aes(seconds)) +
  geom_line(aes(y = Core.Clocks..avg...MHz.)) +
  geom_text(aes(label = "Cinebench Score: 19406"), x = 75, y = 5400) +
  
  labs(
    y = "CPU Speed (MHz)", 
    x = "Time - (Minute:Seconds)", 
    title = "CPU clock Speed During Cinebench Run"
  ) +
  
  scale_x_continuous(
    breaks = seq(0, max(headed_num$seconds), by = 20),
    
    #sprintf() is a function in R that allows you to format strings with placeholders for variables. It takes two arguments: the format string and the variables to substitute into the string. The format string is a character string that includes placeholders for the variables.
    labels = function(x) sprintf("%02d:%02d", floor(x/60), x %% 60)
  ) +
  
  scale_y_continuous(
    breaks = seq(3500, 5500, by = 250),
    limits = c(3500, 5500),
    minor_breaks = seq(3500, 5500, by = 125)) +
  
  #scale_color_manual(name = "Measures",
                     #value = "Core.Clocks..avg...MHz." = "blue") +
  
  theme_clean() +
  
  theme(panel.grid.minor.y = element_line(color = "gray", linetype = "dotted"),
       legend.position = c(0.65, 0.40),
       legend.title = element_text(size = 10.5), 
       legend.text = element_text(size = 10),
       legend.background = element_rect(fill = "grey99", color = "black", size = 0.5))  #makes a box around the legend and adds a fill! 

#The "cairo = FALSE" parameter should get rid of the anti aliasing on the cinebench text
#ggsave(filename = "my_plot.png", dpi = 300, cairo = FALSE)
```


```{r}
#fan speed?
```



###### GPU ######

```{r}
library(tidyverse)
library(readxl)
library(lubridate) #for date extraction and manipulation
library(ggthemes)

hwinfo_cinebench <- read_excel("C:/Users/Ryan/Coding Projects/Computer Experiment/Test Com Data.xlsx", sheet = "hwinfo_cinebench")

#the -2 removes the last 2 observations in the data set which are just descriptors of the variables 
headed <- head(hwinfo_cinebench, -2)

#Converting all the variables in the dataset, expect for date, into numeric variables
headed_num <- as.data.frame(lapply(headed[,-1], as.numeric))

#imputing date variable back into the headed_num dataframe
headed_num <- cbind(headed$`Date/Time`, headed_num)

#Rename the date column we just added back above 
names(headed_num)[names(headed_num) == "headed$`Date/Time`"] <- "DateTime"

# Convert DateTime variable to POSIXct format
headed_num$DateTime <- ymd_hms(headed_num$DateTime)

# Convert DateTime variable to numeric representing seconds since initialization point
headed_num$seconds <- as.numeric(headed_num$DateTime - min(headed_num$DateTime))

#removing unimportant datasets
rm(hwinfo_cinebench, headed)
```


```{r}
#Getting the basic summary statistics for all of the relevant variables in the dataframe 
head_sum <- headed_num %>% select(-DateTime, -seconds)

#Applying the summary function to every variable in our modified head_sum dataframe
summary <- lapply(head_sum, summary)

#Getting the variable names from the summary list so that I can index into specific variable of interst 
head_sum_names_list <- names(summary)

summary[["GPU.Temperature...C."]]

summary
```


```{r}
#Temperature and power consumption graph 
names(headed_num)

headed_num %>% ggplot(aes(seconds, CPU.Package.Power..W.)) +
  geom_line(aes(y = CPU.Package.Power..W., color = "CPU.Package.Power..W.")) +
  geom_line(aes(y = CPU..Tctl.Tdie....C., color = "CPU Tctl Tdie")) +
  
  labs(
    y = "Temperature (*C)",
    x = "Time - Minute:Seconds", 
    title = "Comparing different measures of CPU Temp over time"
  ) +
  
  scale_x_continuous(
    breaks = seq(0, max(headed_num$seconds), by = 20),
    
    #sprintf() is a function in R that allows you to format strings with placeholders for variables. It takes two arguments: the format string and the variables to substitute into the string. The format string is a character string that includes placeholders for the variables.
    
    labels = function(x) sprintf("%02d:%02d", floor(x/60), x %% 60)
  ) +
  
  scale_y_continuous(name = "Watts" ,
    sec.axis = sec_axis(~., name =  "Temperature (*C)")) +
    #breaks = seq(40, 80, by = 5),
    #minor_breaks = seq(40, 80, by = 2.5),
    #limits = c(40, 80)) +
  
  scale_color_manual(name = "CPU Temp Measures",
                     values = c("CPU Tctl Tdie" = "red", "CPU.Package.Power..W." = "blue")) +
  
  theme_bw() +
  
  theme(panel.grid.minor.y = element_line(color = "gray", linetype = "dotted"),
       legend.position = c(0.65, 0.40),
       legend.title = element_text(size = 10.5), 
       legend.text = element_text(size = 10),
       legend.background = element_rect(fill = "grey99", color = "black", size = 0.5)) + #makes a box around the legend and adds a fill! 
  
  guides(fill = guide_legend(reverse = TRUE)) 
```



#Useful code for the future:
scale_x_datetime(date_breaks = "5 sec", date_labels = "%M:%S") - formats datetime in a specific way within ggplot
scale_x_discrete(breaks = c()) - eliminates x axis labels 


########################### FrameView ##############################################


########## Synthetic ##########

#Frametime plots + other useful statistics from GPU and CPU (no GPU power and CPU voltage numbers though)

```{r}
Frametime_GTAV <- read_excel("C:/Users/Ryan/Coding Projects/Computer Experiment/Test Com Data.xlsx", sheet = "frameview_gtav")
Frametime_ow <- read_excel("C:/Users/Ryan/Coding Projects/Computer Experiment/Test Com Data.xlsx", sheet = "frameview_ow")
Frametime_R6 <- read_excel("C:/Users/Ryan/Coding Projects/Computer Experiment/Test Com Data.xlsx", sheet = "frametime_r6")
Frameview_Apex <- read_excel("C:/Users/Ryan/Coding Projects/Computer Experiment/Test Com Data.xlsx", sheet = "frameview_apex")
Frameview_Destiny <- read_excel("C:/Users/Ryan/Coding Projects/Computer Experiment/Test Com Data.xlsx", sheet = "frameview_destiny")

he <- Frametime_R6 %>% select(TimeInSeconds, MsBetweenPresents, MsBetweenDisplayChange, `GPU1Clk(MHz)`, `GPU1Util(%)`, `GPU1Temp(C)`, `AMDPwr(W) (API)`, `CPUClk(MHz)`,	`CPUUtil(%)`, `CPU Package Temp(C)`, `CPU Package Power(W)`)

he$Seconds <- as.numeric(he$TimeInSeconds - min(he$TimeInSeconds))

#Way to thin out the observations in my dataset, although it makes no difference at large time intervals, once we zoom in on a specific time thinning out the data helps to smooth the time series graph and avoids it being so jagged  
#This specific code reports values of variables 3 times a second 
df_subset <- he[seq(1, nrow(he), by = 50), ]
```

#Frametime Graph - need to be careful plotting frametimes because more polls per second may check more issues

```{r}

he %>% ggplot(aes(Seconds, MsBetweenPresents)) + 
  geom_line(color = "blue") +
  labs(
    y = "Frametime",
    title = "Frametimes"
  ) +
  
  #scale_x_continuous(breaks = seq(0, 600, by = 100), 
                     #limits = c(0, 600),
                     #minor_breaks = seq(0, 600, by = 50)) +
  
  theme_bw()


#subsetted data doesn't catch the spikes in frame time that lead to bad performance that the original dataset catches 
#df_subset %>% ggplot(aes(Seconds, MsBetweenPresents)) + geom_line() #+ #scale_x_continuous(breaks = seq(0, 30, by = 10), limits = c(0, 30)) 
```


#graphing specific varaibles over time from the FrameView dataset 
```{r}
names(df_subset)

df_subset %>% ggplot(aes(Seconds, `GPU1Temp(C)`)) +
  geom_line() +
  
  labs(
    title = "CPU Power Consumption Over Time",
    x = "Time - Minute:Seconds"
  ) +
  
  scale_x_continuous(
    limits = c(0, 90),
    
    #sprintf() is a function in R that allows you to format strings with placeholders for variables. It takes two arguments: the format string and the variables to substitute into the string. The format string is a character string that includes placeholders for the variables.
    
    labels = function(x) sprintf("%02d:%02d", floor(x/60), x %% 60)
  ) +

   scale_y_continuous(breaks = seq(0,60, by = 15),
                     limits = c(0,60)) +
  
  theme_bw() 
  
```


#Code to animate a line graph to plot over time
```{r eval=FALSE, include=FALSE}
library(ggplot2)
library(gganimate)

# Create data for animation
anim_data <- data.frame(
  Seconds = seq(0, 600, length.out = 601),
  MsBetweenPresents = rnorm(601, mean = 16.67, sd = 3),
  frame = seq(1, 601)
)

# Create base ggplot object
base_plot <- ggplot(anim_data, aes(Seconds, MsBetweenPresents)) + 
  geom_line(color = "blue") +
  labs(
    y = "Frametime",
    title = "Frametimes"
  ) +
  scale_x_continuous(breaks = seq(0, 600, by = 100), 
                     limits = c(0, 600),
                     minor_breaks = seq(0, 600, by = 50)) +
  theme_bw()

# Add animation to plot
animated_plot <- base_plot + 
  transition_reveal(frame) +
  view_follow(fixed_x = TRUE)

# Render animation
animate(animated_plot)

library(magick)

# Define the file names and ordering
file_names <- paste0("gganim_plot", sprintf("%04d", 1:100), ".png")

# Read in the pictures
imgs <- image_read(file_names)

# Combine the pictures into an animated GIF
animation <- image_animate(imgs, fps = 10)

# Save the animation to a file
image_write(animation, "animation.gif")
```



########## Gaming ##########
 

#Averge FPS, 1%, and .1% lows for game tests 

```{r}
#Data Wrangling 

summary_fps <- read_excel("C:/Users/Ryan/Coding Projects/Computer Experiment/Test Com Data.xlsx", sheet = "summary")

#contains columns I might find important later but don't rn like `Min FPS`, `Max FPS`
summary_fps_o <- summary_fps %>% select(Application, `Avg FPS`, `1% Low FPS`, `0.1% Low FPS`, `Min FPS`, `Max FPS`)

summary_fps <- summary_fps %>% select(Application, `Avg FPS`, `1% Low FPS`, `0.1% Low FPS`)

#code to strip the .exe from the application variable so that it just displays the game name 
pattern <- "\\.exe$"

summary_fps$Application <- str_replace(summary_fps$Application, pattern, "")

hehe <- pivot_longer(summary_fps, cols = c("Avg FPS", "1% Low FPS", "0.1% Low FPS"), names_to = "fps_type", values_to = "fps")

#changing the format of the newly created FPS column so that it only displays out to 1 decimal point string to display all numbers 
hehe$fps <- round(hehe$fps, digits = 1)
```


```{r}
# Plot horizontal bar chart for all selected games 

hehe %>% ggplot(aes(x = fps, y = Application, fill = fps_type)) +
  geom_text(aes(label = fps), hjust = -0.1, position = position_dodge(width = 1)) +
  geom_bar(stat = "identity", position = "dodge") +
  
  labs(
    x = "FPS", 
    y = "Game", 
    fill = "",
    title = "Gaming Benchmarks March 2023",
    subtitle = "See spec sheet for specific computer configuration") +
  
  scale_fill_manual(values = c("Avg FPS" = "darkblue", "1% Low FPS" = "maroon", "0.1% Low FPS" = "yellow")) +
  
  #scale_x_continuous(breaks = seq(0, 200, by = 20),
    #limits = c(0, 200)) +
  #Reversing the legend does work in this case, but like in the KDL study I got around this by manually defining the factor levels for each variable 
  guides(fill = guide_legend(reverse = TRUE)) +
  theme_bw()
```