Assignment 2

Minicase

Background

You work for BostonRealty, and, for the purposes of this minicase, you will be playing two roles:

1) Business Development Manager

2) Member of the Data Science Team

The VP of Strategy has outlined a strategy of expanding for the company. This involves bidding on properties in Boston that do not have assessed values, and the VP of Strategy has tasked you to design a framework that will provide guidance on what properties to bid for.

In your role as Business Development Manager, you ask your data science team to build a First Cut Predictive Model. You also tell the team that you do not care about how "accurate" or sophisticated this model is. You just want the model to be supported by good reasoning and appropriate analysis. You also want to get the project off the ground quickly and once they have the First Cut Predictive Model ready, you will work with the team to refine it.

In your role as Member of the Data Science Team, you do the following:

1) You obtain a data set for the housing market in Boston's West Roxbury neighborhood. Note: This dataset is called WEST ROXBURY DATA SET ONLY.

2) You follow all the steps to build a First Cut Predictive Model. Note: These steps were reviewed in a video by Thomas.

Once the First Cut Predictive Model is done, you put on your Business Development Manager hat and ask your data science team the following questions:

1) Did you uncover any anomalies in your dataset? If yes, what were the specific anomalies (be sure they are different from the ones Thomas pointed out.) Why are they anomalies? How did you specifically address them?

2) From your First Cut Predictive Model are there coefficient estimates whose signs do not seem to make sense from a theoretical or business perspective? Explain why those signs do not make sense.

3) What are both (1) the direction (under-predicting or over-predicting) and (2) magnitude of the forecast errors from this First Cut Predictive Model?

Task

Your task, as a member of the data science team, is to fill out the Assignment 2 Response Table below (in Microsoft Word) that addresses the 3 questions posed by the Business Development Manager:

1) What are the specific anomalies you uncovered in your dataset? Note: Be sure they are different from the ones Thomas pointed out. Why are they anomalies? How did you specifically address them? NOTE: be specific and precise.

2) From your First Cut Predictive Model, what coefficient estimates have signs that do not seem to make sense from a theoretical or business perspective? Explain why those signs do not make sense.

3) What are both (1) the direction (under-predicting or over-predicting) and (2) magnitude of the forecast errors from this First Cut Predictive Model?