Stress detection and activity peaks

This project is a statistical and exploratory data analysis approach to measure stress and physical activity through data provided by smart watches.

To this end, two main analyses were conducted to understand the data and to find a statistical approach to measure and evaluate health data. An individual study and a group comparison.

Data Mining

The study contains 8 datasets from different volunteers who have exported their data from the iphone Health App (Apple Health) for this project.

Tech stacking

• Pyhton programming language
• Libraries:
  🐼 Pandas
  📊 Seaborn
  📈 Matplotlib
  📇 Numpy
  📦 Pickle
  🗓️ Datetime
  🧮 Scipy

With the help of data manipulation techniques, statistics, exploratory data analysis, data cleaning and visulization, we were able to observe peaks of stress indvidually and comparisons of physical activity for the study subjects.

To achieve this we have made use of pyhton, applying mainly the pandas, seaborn and matplotlib libraries together with libraries focused on calculation and statistical analysis such as scipy and numpy.

Index

The Data folder contains the original datasets used in the project.

In the Notebooks folder are the different files that were used for the Python work:

• Cleaning_func.py: File with the functions used for data cleaning.
• Transform_func.py`: File with the functions used for data transformation and organization.
• Cleaning-Transforming-Final.ipynb: Notebook with the complete steps for the Data Wrangling.
• Statistical-analysis.ipynb: Notebook with the individual analysis of one of the participants in the study to identify stress peaks and their respective conclusions.
• Statistical-analysis_2.ipynb`: Notebook with the comparative analysis for all individuals in the study in terms of their physical activity patterns, and their respective conclusions.

Finally, the file EDA Presentation.pdf contains the supporting slides for the presentation of this project.