Summary
This project was about learning to manipulate big data. To do this, my class was split into groups. Each group had to design an experiment to test subject's walking patterns, and record data on them. Then, every group had to create a model from the data and use it to predict a gait trait of new subjects.
My group used an accelerometer to measure g-force of a subject in all three spacial dimensions, and also recorded height, leg height, weight, and gender. After collecting data points on seven initial test subjects, we found a correlation in our data between height and "stepping force." (Stepping force was recorded as g-force measured in the y direction with respect to the ground.) We then tested this model on an additional three subjects, and found we could predict someone's height based on their stepping force with around 3% inaccuracy.
My group used an accelerometer to measure g-force of a subject in all three spacial dimensions, and also recorded height, leg height, weight, and gender. After collecting data points on seven initial test subjects, we found a correlation in our data between height and "stepping force." (Stepping force was recorded as g-force measured in the y direction with respect to the ground.) We then tested this model on an additional three subjects, and found we could predict someone's height based on their stepping force with around 3% inaccuracy.
Key Vocabulary
Gait: A person's manner of walking. Gait was the main thing every group had to collect data on.
G-force: A force acting on a body as a result of acceleration or gravity. 1 g of force is equal to the weight of the body. My group used g-force as a variable correlated to height in our predictive model. Coefficient of determination: The proportion of the variance in the dependent variable that is predictable from the independent variable. The coefficient of determination for our model was around 0.74. This means we could explain 75% of the variance in height of all of our test subjects by just one variable: g-force. |
|
Reflection
Overall, this project was a success. One peak I had was data manipulation. I found the process of using graphs and equations to analyze thousands of data points to be really interesting. It allowed my group to notice macro trends and patterns in our data which would have otherwise been impossible due to the vast amount of numbers to comb through by hand. Another thing I believe went well was our presentation. We finished testing our model ahead of schedule, which left us room to practice our presentation in order to streamline our slides and be able to expand on key points.
One area I felt my group was lacking was collaboration. One member was often not present, which led to communication being disrupted. This made it difficult to keep everyone up to date on our progress, which slowed down our model's development.
One area I felt my group was lacking was collaboration. One member was often not present, which led to communication being disrupted. This made it difficult to keep everyone up to date on our progress, which slowed down our model's development.