For this project you will examine a set of data that you can apply the regression model to.

Process

Step 1

You need to find a set of data, preferably related to but not restricted to your field of study. The data need to fit the regression framework with a regression variable x and a response variable y. Please carefully make the assignment as to which variable is xand which y. It may be necessary to consult a journal or periodical from your field if you do not have other research data available.

Example Data Set

Some of you may not have a data set available to you. If not, feel free to search out a data set that will fit the regression framework. One example data set that can be used comes from the Weather Underground, which collects historical weather data for 30 cities. If you would like to use this data, go to the Weather Underground (Links to an external site.) website. Click on More and from the drop-down menu choose Historical Weather. On this page, you can enter any city and date.

Next, create a well-labeled scatter-plot with the difference of mean temperature and dew point (difference = mean temperature - dew point) on the x-axis and average humidity on the y-axis. You’ll likely need to create a new column for difference in your spreadsheet. It works best to create a new column in Excel for the difference of mean temperature and dew point. Comment on what you see.

Think about if linear correlation and regression suitable tools to describe this relationship?

Step 2

Plot y versus x. Comment on the relationship as seen from the plot.
NOTE Please use Excel or StatCruch to plot this relationship.

Step 3

Fit an appropriate regression model from the data. Use simple linear regression or fit a polynomial model to the data. Comment on measures of quality.

Step 4

Plot residuals as illustrated in the text. Check possible violation of assumptions. Show graphically a plot of confidence intervals on a mean response plotted against x and comment.