Move Yourself-Data Analytics Teaching App (MY-DATA)

Introduction

MY-DATA is an Android-based mobile app for learning data science. MY-DATA allows users to capture their own videos using the built-in camera on their mobile device. By detecting key joint points (e.g., shoulder, knee) in these videos, MY-DATA creates personalized motion data sets from which students can learn data science.

To enhance students’ learning experience, MY-DATA provides strong data visualization, including a variety of visual aids such as interactive graphs and figures, that they cannot otherwise experience in traditional classroom settings. Students can enjoy a customized learning experience with engaging visual cues at their own pace.

MY-DATA currently offers two learning modules dedicated to learning principal component analysis (PCA) and k-means clustering. To ensure students have a well-rounded understanding of data science, new learning modules addressing the diverse methods in data science will be introduced.

SO STAY TUNED WITH US 😃

Overview

With MY-DATA, you’ll go through a four-stage learning process🚀:

• Stage 1. method selection

• Stage 2. brief method introduction

• Stage 3. data collection through self-body motion

• Stage 4. method learning with the collected data.

Stage 1. Method Selection

To pick a topic you wanna learn, simply tap the button Dimensionality Reduction or Clustering. It’s that easy! In the future, as we add new learning modules, you’ll be able to scroll horizontally through the buttons. Easy-peasy!😎

Stage 2. Method Introduction

As warm-up🚶🏿‍♂️, MY-DATA will make sure you have all the necessary background knowledge to understand the method. This will help you hit the ground running!🏃🏿‍♂️💨

Stage 3. Full Body Motion Data Collection

To create your own one-of-a-kind body motion dataset, all you need is just one tap.🎉 MY-DATA will capture you while showing your 12 key joint points in real-time for 10 seconds.

PCA Learning Module

Get ready to learn PCA! In this module, you’ll extract principal components (PCs) from your walking motion dataset and visualize your walking pattern through PC1 and PC2. Don’t miss the chance to be amazed😮 by these two features on the scree plot!

• K-means Clustering Learning Module

Now, it’s classification time! In this module, you’ll use the PCA technique you learned earlier to classify your walking, standing, jumping jack motions with data visualization. You’ll love seeing how k-means clustering improves💪 as you find better and better centroids!

Open Source

• All the source code and resources are available at our GITHUB repo.
• You can download the .apk file here to test MY-DATA on your device.
• Have questions? Just ask! And if you notice any errors, feel free to reach out to me at sjung9@ncsu.edu.

For More Information

For more information or to view similar projects, please visit the Biomechanics publications page.

Citation

• Please cite the following paper if you use MY-DATA or the code above in your research or development.
  • Jung, S., Wang, H., Su, B., Lu, L., Qing, L., Fang, X., & Xu, X. (2024). Learning Undergraduate Data Science Through a Mobile Device and Full Body Movements. TechTrends, 1-13.