Plot Linear Regression Model
- Residuals -vs- Fitted
- Noraml Q-Q
- sqrt of Standardized Residuals -vs- Fitted
- Standardized Residuals -vs- Leverage
jay Yanamandala
11/5/2021
This presentation is related to predicting home prices in King County Washington State. Since the final assignment is related to Developing Data Products CourseRA course, to keep it simple, prediction is based only on:
1. Sqft Living
2. Number of Bedrooms
3. Number of Bathrooms
This reproducible pitch and Shiny App are presented to showcase the prediction model
Access the shiny app here
Access the github repo that hosts the files here
The dataset was provided for one of the assignments in “Machine Learning Specialization” offered by Univ of Washington, and taught by:
– Emily Fox, Amazon Professor of Machine Learning
– Carlos Guestrin, Amazon Professor of Machine Learning
Before defining Shiny ui.R inputs, we clean the dataset for simple presentation
* Remove columns that are not needed for the current analysis
* Convert integer to numeric
* Update columns that have value ‘0’ with mean of previous five values
* Sort data for sliders in ui.R
Printing first 10 rows, and few columns of dataset
bedrooms bathrooms sqft_living yr_built zipcode
1 3 1.00 1180 1955 98178
2 3 2.25 2570 1951 98125
3 2 1.00 770 1933 98028
4 4 3.00 1960 1965 98136
5 3 2.00 1680 1987 98074
6 4 4.50 5420 2001 98053
7 3 2.25 1715 1995 98003
8 3 1.50 1060 1963 98198
9 3 1.00 1780 1960 98146
10 3 2.50 1890 2003 98038In our ui.R file we define the following inputs
* slidersqft - Sqft of living space
* sliderbed - Number of Bed rooms
* sliderbath - Number of Bath rooms
* showModel - Radio button to select which ‘lm’ model to plot
In our server.R we capture the input from ui.R and
* Plot a model - default plot is sqft_living + bedrooms + bathrooms
* Print a table of 3 models
1. price -vs- sqft_living
2. price -vs- sqft_living + bedrooms
2. price -vs- sqft_living + bedrooms + bathrooms
Value of Predicted home sqft 2750, bedrooms 5, bathrooms 3.5 is:
$663,160.9764