1. (4 pts) The le \face-data.txt" contains the pixel values of four 64 64 images of faces, captured from di erent angles. For each image, the pixel values are stored in one column, in the order of vectorizing a 64 64 matrix (4096 pixel values). So, the data are stored as a 4096 4 matrix. Visualizing each of the 64 64 matrix will give one such image (actually, this is one of the most basic ways to store an image digitally). Visualize the four images in R using the matrix visualization methods covered in the lecture. Potentially, you may need some rotation or ipping to show the images in the correct orientation (clearly, it does not make sense if you have them upside down). Show the four faces in a 4 1 layout. You can pick up whatever coloring theme you want. (2 pt for correctly show each all the faces with clearness, 1 pt for correct orientation, 1 pt for correct layout). We will also use this data set for later sections.
2. (6 pts) An older teacher (in her 40’s) applied for a job in her local school system. She was not hired and did not even receive an interview. She claimed to be the victim of age discrimination and believed that the school preferred to hire younger teachers. If true, this would be against the law in the United States. Her lawyers obtained a data set \TeacherHires.csv" on applicants for teaching positions from the school les. You are supposed to do exploratory visualization to analyze the data and give certain insights into the problem. Although this has been partly processed, it is still very messy. I would do some necessary cleaning and exploration with you in class. However, you will need to clean it further to ensure it is usable for your own analysis. Conduct an exploratory analysis: is there evidence of age discrimination in the interviewing and/or hiring of teachers? Use your visualization analysis to answer this question. It is an open problem, clearly. Give your answer and justify your conclusion with your
1
visualization analysis. (2 pts for the basic presentation of the facts from the data, 2 pts for e ective and clear visualization, 2 pts for insightful and convincing analysis.)
For those of you who have di culty starting the analysis, one suggestion is: start with asking simple and concrete questions, then gure out how to answer that question, and based on it, raise your next question with your nal target in mind. You can explore any variables in the dataset that interest you or that might be important. For example, if hired teachers have higher GPA scores, this could be important in your analysis. If you prefer to treat age as \continuous", that is ne, too. You do not have to include your data cleaning step in your submission, but it will be crucial to ensure your work quality.
2
