Description

Open the file HousingBubble.xlxs.  It contains data with respect to house features, assessments and selling prices, before the 2008 housing bubble and afterwards.  In 2008, there was a collapse of the housing industry in the USA that triggered a financial crisis that spread around the world.  People could no longer pay their mortgages.  As a result, there was a surplus of housing that was for sale and prices became depressed.  I do not know the exact source of this dataset.  It was provided by a US publisher.  I believe that the data is authentic.

For each house we are given data with respect to the size of the home, the number of bedrooms, bathrooms and other rooms, and the size of the deck or garage, if there is one.  We are given the age of the home as well as an indicator of whether the house is in particularly good or poor condition.  Each property has an assessment of the value of the land and of the building.  An assessment is determined by the local town as the basis for assessing property taxes.  It reflects the estimate of what the property should sell for if offered for sale.  We are also given the selling price.  The selling price is what the property actually sold for, whereas the assessment is what the town thought it was worth.

How does the town assess property?  This is a value estimation problem.  One could build a mathematical model that incorporates known property features to forecast selling price.  This model could then be used to determine assessments for unsold properties.  Unfortunately, it is challenging to build a model that includes location and qualitative features of a property.  Location is a critical factor in housing prices.

Nonetheless, this data set provides some useful illustrations in building and interpreting regression models.  We will look only at the data Pre-Crisis in this assignment.

Excel requires that variables selected for Correlation or X variables used in Regression must be adjacent (beside) one another.  For the questions below, in many cases you will be selecting a set of variables that are not clustered together.  I strongly recommend that for each question, you select the variables that you will need and then copy them to a new sheet.  Have a separate sheet for each question.  You can rename sheets 1, 2, 3a,… so that you can keep track of which sheet is for which question.

1.       How good are the assessments as indicators of house prices?  For the Pre-Crisis data set of 1977 homes, the average selling price was $116,208 and the average assessment was $114,985, so, on average, the assessments slightly underestimated the price.  But were the assessments good indicators of price?

a.       Construct a scatter chart of price versus assessment, with price on the vertical axis.  Make sure you label your axes.  Take a screenshot of your chart and paste it into your assignment.

b.      Calculate the correlation between price and assessment using the Correlation function in the Data Analysis tools.

c.       Do you think the assessments are good predictors of price?