Individual Paper - I BUSI 650 – Business Analytics The paper must include the following, · Discuss data strategy of a company of your choice. (~700 words) · Create a regression model with at least 20...

Introduction
In this project we are building a predictive housing price model with maximum accuracy using linear regression. First we want to find a list of significant contributors to house prices. We have thirteen independent variables. We have done feature engineering to clean the data. We have chosen a linear regression estimate to predict the house price of Boston data. We have eliminated insignificant predictors from the model. Some variables were eliminated inorder to avoid any kind of multicollinearity presence. Multicollinearity severely affects our model performance and predictive ability. We have partitioned data into two group validation and training. In future several other machine learning models can be applied to this data in order to boost efficiency of the model.
Importance of the chosen area:
Due to exponential growth in the real estate market whether it's the rental market or property market. We have witnessed the success of AirBNB, VRBO and other housing services providers. They have built models to evaluate property prices depending on many significant factors. Housing platforms could be another reason to build this model.
We can build a prototype using a web page on which we enter several essential information regarding the house like location, owner origin and income. We will train models with existing data. We will use this trained model to help the end user predict house prices of their property. Whenever users or clientele enter certain required information they could get an estimated price of their house.
We have to gather as much data as possible to build a more accurate trained model. More relevant features or variables will eventually increase the explainability of the model. We can also run diagnostic analysis to find main attributes behind increasing prices of houses. What are those attributes? How to make informed decisions about houses? In every model we have a certain amount of uncertainty present or in classification models false positive and false negative percentages. We have to minimize those uncertainties in order to achieve reliable accuracy. In our model we will remove all insignificant variables and remove significant variables that are causing multicollinearity. which will adversely affect our model performance.
Another ubiquitous application would be to build a portal to entry in the short term rental market. Many world renowned companies use advanced analytics to optimize long term rental market rates and occupancy rates in order to maximize profits. We could use sensitivity analysis to find tradeoffs between occupancy and rental rates. As we know rental rates and occupancy rates are inversely proportional to each other.
Why this data set is interesting
This data was first used in the 1976; journal of environmental economics and management. We can use this data to build a predictive model for house prices. That can be used for house rental purposes to better evaluate house prices or real estate properties sales platform or entity.
What has been done so far:
Earlier researchers have investigated problems related with housing price data to measure the respondent willingness to pay for clean air. They have used hedonic price models and Boston housing data. They have calculated estimates of respondent willingness to pay for clean air.
We are using this data to predict prices using boston data variables,for example crime rate in the area. We can build a linear regression model to estimate house prices for many applications like rental Market...