smartphone_sensor_predictive_modeling

Folder Structure:
requirements.txt
scripts
    clustering.py
    __init__.py
    __pycache__
    visualization.py
    preprocessing.py
    modeling.py
notebooks
    01_preprocessing.ipynb
    02_clustering.ipynb
    03_modeling.ipynb
    04_results.ipynb



Steps of process:
1. Preprocessing
  - merging datasets together
  - imputation
  - encoding missingness at certain thresholds
2. Exploratory Data Analysis
  - Plotting number of participants for each week
  - Plotting number of participants for each week for each variable
  - Feature selection with linear regression
  - EDA: plotting target variable (depression score) against single selected features
3. Clustering
  - PCA of features -> groups of modality (texts, calls, mobility, activity, etc)
  - PCA of subjects' symptom trajectories
    - by depression score over time
    - by mobility over time
    - by activity over time
  - Clustering of participants into demographic groups
4. Cluster Analysis
  - Bar plots of various clusters for target variables
  - Plotting graphs and descriptive statistics for clusters vs. target variable over time
5. Preprocessing for machine learning
  - Scaling
  - Encoding categorical variables
  - Creating wide dataframes
  - Creating lagged dataframes
5. Modeling Linear Regression and Mixed Effects Linear Regression
6. Modeling Decision Trees
7. Modeling Random Forest