Eat & Tell: A Randomized Trial of Random-Loss Incentive to Increase Dietary Self-Tracking Compliance

This repository contains the Jupyter notebook used for reproducing the results published in our Digital Health 2018 paper:

Achananuparp, P., Lim, E.-P., Abhishek, V., & Yun, T. (2018). Eat & Tell: A Randomized Trial of Random-Loss Incentive to Increase Dietary Self-Tracking Compliance. In Proceedings of the 2018 International Conference on Digital Health - DH ’18. https://doi.org/10.1145/3194658.3194662

Please contact Aek if you have any questions or problems.

Requirements

The notebooks have been tested in R 3.5.1 via Anaconda with the following packages:

• effects
• dlyr
• ggplot2
• lme4

Project Structure

By default, the project assumes the following directory structure:

project 
└───data  
│   │   deduction_amounts.csv
│   │   deductions_7d.csv
│   │   deductions.csv
│   │   demos.csv
│   │   end-of-days_7d.csv
│   │   end-of-days.csv
│   │   food_diaries.csv
│   │   post_food_diaries.csv
│   │   pre_food_diaries.csv
│   │   users.csv
└───notebooks
│   │   data-analysis.ipynb
└───reports
│   └───figures

All CSV data files should be put in the data folder. All notebooks should be put in the notebooks folder. Any generated reports and figures will be put in the reports folder.

Pipeline

Step 0: Data import

Download the data and extract the CSV files to the data directory.

Step 1: Run the data analysis notebook

Run the notebook data-analysis.ipynb to perform all analyses.

Outputs: Several figures will be generated and stored in the reports/figures folder.