Module 8: Linear Regression Visualizations in ggplot2

 Module 8: Linear Regression Visualizations in ggplot2

DISCLAIMER: In this Week's submission, I will post the code and my comments here, but will submit a separate HTML .rmd file for your convenience in Canvas, so that viewing my work is easier since it will be organized into a compiled document.

In this weeks module, we explored different styles of Statistical Analysis from a Dataset in R programming. The key, is to remember when dealing with data, what is the appropriate time to separate (use factor()) or to treat data irrespective of each other by allowing unfiltered data directly from the source to be visually represented. I used related columns from the data frame to visualize similarities and trends amongst characteristics of automobiles that I know for a fact can tell us about the others even without knowledge of their data. Without data, one might assume a larger engine means a quicker quarter mile or less fuel mileage, but that is not always the case. This data does not take into account possible modifications or other nuances of the car from the factory that might make it an outlier in its respective category. Instances like this are what make for small anomalies in the data, but from what was provided to me in the dataset, here are some visualizations on what I believe to be an accurate regression analysis in ggplot2.

##Module 8: Linear Regression Analysis in ggplot2

## Let's conduct a simple regression analysis in R using mtcars

##First, let's call 'mtcars' and set it to CD

##Add Some libraries too



library(dplyr)

library(tidyverse)

library(ggplot2)



data(mtcars)

CD <- data.frame(mtcars)


#Now let's isolate some data and generate a quick histogram 

#Say you want to analyze the relationship between MPG and the Number of Cylinders 

#First, define ggplot(), the data and then the aesthetics (aes()), and then specify some colors and labels

## To add a regression line or line of best fit, use geom_smooth.



ggplot(CD, aes(CD$wt,CD$mpg)) +

  geom_point(

    fill = "blue",

    color = "black"

  ) +

  labs(

    x = "Cylinder Count",

    y = "MPG"

  ) +

  geom_smooth(method = "lm")



#You can also use stat_smooth() to achieve the same effect as geom_smooth()


ggplot(CD, aes(CD$wt,CD$mpg)) +

  geom_point(

    fill = "blue",

    color = "black"

  ) +

  labs(

    x = "Cylinder Count",

    y = "MPG"

  ) +

  stat_smooth(method = "lm", color = "red")


#I changed the color to red to specify the difference in the same line being designated by two different function calls

## Let's try something a little different. Instead of using a scatterplot and regression line

##We can do a line of best fit for each varable, see the scenario below:


#Let's visualize a regression analysis line on a scatterplot between, engine power (hp) and quarter mile times (qsec).

## We will set the color to the cylinder number to show that some larger engines that use more fuel may in fact be faster


#I make another object with my desired title to avoid using factor() in more that one place or within the title

##Call it, my way of keeping my presentation and visualizations clean.

Cylinders <- factor(CD$cyl)


ggplot(CD, aes(CD$hp,CD$qsec)) +

  geom_point(

    fill = "blue",

    color = CD$cyl

  ) +

  labs(

    x = "Engine Output (HP)",

    y = "Quarter Mile Time (qsec)"

  ) +

  stat_smooth(

    method = "lm",

    se=FALSE,

    fullrange=TRUE,

    aes(color= Cylinders))


#We can also make use of a grid, within the theme() function call


ggplot(CD, aes(CD$hp,CD$qsec)) +

  geom_point(

    fill = "blue",

    color = CD$cyl

  ) +

  theme(panel.grid = element_line(color = "#8ccde3",

                                  size = 0.75,

                                  linetype = 2)) +

  labs(

    x = "Engine Output (HP)",

    y = "Quarter Mile Time (qsec)"

  ) +

  stat_smooth(

    method = "lm",

    se=FALSE,

    fullrange=TRUE,

    aes(color= Cylinders))



#As you can see from the resulting visualization, the correlation between a quarter mile time is predecated on a few variables

##First, as the number of horsepower increases, the populated points read higher hp, are larger cylinder count engines, and by method of deduction, have a larger displacement

##The results show that as you increase the number of cylinders and the horsepower, your quarter mile time is going to go down.

##Though, one might believe that 8 cylinders should out perform 6 cylinders, but I must remind that some of the cars on the list in mtcars, are 6 cylinder

##but have either individual throttle bodies or forced induction which creates more power and pickup in a smaller engine thats lighter and grips the ground easier.

##Ultimately you would have to test that individually or add that data to mtcars, if the car has a turbo/supercharger or not and if so? how many PSI of boost does it give at maximum load capacity?

##Other than that, this is the end of this weeks review of Linear and Multiple Regression Analysis in ggplot2.

Comments

Post a Comment

Popular posts from this blog

R Package: pfStat

Image
 Final Project: pfStat, a personal finance mini-suite in R GitHub URL: https://github.com/tahousex/pfStat What is pfStat? This package is entitled pfStat, and is a small package with a few functions dedicated to helping a user calculate different personal finance metrics. The functions include the ability to derive Returns on Interest, Tax Rates based on Income, Taxable Income and Home Equity, and the purpose behind choosing these kinds of metrics was to emphasize the importance of knowing where one puts their wallet. It goes without being entirely sanctimonious that the dollar does not go as far as it once did, and knowing where one stands amongst their own finances and parameters in the financial world will assist them in making more calculated decisions.  The Functions: The taxrate() income receives input from the user income and state tax percentage represented as a decimal percentage (0-1)   and gives the user output of a dollar amount of the taxes owed based on the input The taxa
Read more

Module 7 Assignment

Module 7 Assignment  R Code: # Module 7 Assignment #1.1 #First define the data set required, I/O values x <- c(16, 17, 13, 18, 12, 14, 19, 11, 11, 10) y <- c(63, 81, 56, 91, 47, 57, 76, 72, 62, 48) #Then we must create the relationship (correlation) cor(x,y) #This value 0.728 that is in the console tells us a few things #This relationship is correlated and is increasing linearly #I may choose to plot these findings using plot() plot(x,y) #See the lower righthand window for a composite plot for this data, as the regression trednline would positively increase through the points #1.2 #This settion of code will get the value for the coefficients of the linear regression for this data Regression_model=lm(y~x) Regression_model$coefficients #2.1 #This problem will go over the aquisition of similar values with a different data set discharge <- c(3.600,1.800,3.333,2.283,4.533,2.883) waiting <- c(79,54,74,62,85,55) plot(discharge,waiting) #2.1 Regression_model_1=lm(waiting~discharge)
Read more

The Tampa Feasibility Report featuring R based Visualizations

Image
  Visualizations for Professional Writing's Final Deliverable These visualizations were taken from FDOT data for our local district on the subject of transit usage, and crash data that was munged into relative ratios that had a regression performed on these variables within.  Pie Chart The above figure is a pie chart using data collected that estimates the higher end limits of hourly numbers of passengers that could utilize each method of public transportation during peak hours Transit Regression This figure here shows the monthly totals for crashes, fatalities and serious injuries from accidents recorded within FDOT data and the regression line method was used against the 'free_y' inputs which allowed there to be a 95 percent confidence regression on the other two variables from the one in question showing certain trends within and between variables. Bus versus Pubic Transport with 95 percent regression line fitted The same style of regression was sued here but this time i
Read more