Machine Learning (II)


A couple weeks ago, each of you participated in data collection and now the data is ready to analyze.

raw_touch_data.xls.zip

Here are some important columns:

  • TIME (ms) - ms since the beginning of the session
  • STUDENT ID - unique identifier associated with a student
  • TOUCH ID - identifier associted with a touch event (each action event generated multiple sensor readings)
  • TOUCH LABEL - 0 means left, 1 means right

All the other columns are from various sensors.

Challenge

Challenge 1

The first challenge is to use Tableau to answer some basic questions. For each question, take a screenshot of the entire Tableau interface and submit.

a. What is the distribution of the Pitch (Degrees) values?

pitch

b. What do the sensor reading patterns differen across individual students in terms of Pitch (Degrees)?

pitch_students

c. Is there a significant difference between the two touch types (TOUCH LABEL) in terms of the average Pitch (Degrees)?

avg_pitch

d. How many unique touch events did students have?

student_touch

Challenge 2

The second challenge is open ended. Write your own analysis of this data. Keep the analysis short (300 words max). Include visualizations to support your analysis.

(As a heads up, we will do some simple classifier training on this data next week).