seaco_smartwatch_paper

Malaysia SEACO-CH20 Smartwatch Feasibility Study

Data

First thing to do is to get access to the SEACO RDSF - ask the PI for this. Then you'll need to mount it somewhere - the mount location that I used is in userconf.yml.

Environment

Run conda create -f environment.yml to create an environment that contains all the required packages, including R, python and the required libraries.

If you have trouble with this, try installing R and python separately.

Notebooks

Approximately correspond to the order of things in the paper:

• demographic_summary.ipynb: stats on the participant demographics
• survey.ipynb: quantitative results from the survey
• meal_stats.ipynb: statistics on the numbers of meals, snacks, etc. per day
• three_level_model.ipynb: linear models for response rate as the study progesses

Linear Models

If you're just interested in the linear model: three_level_models.R.

The notebook three_level_model.ipynb creates a .csv file for boolean response/not for each hour/day/participant. Then runs the models in an R subprocess, then calculates odds ratios based on the log odds in the files output from the R scripts (which ive copied over manually).

There are also additional notebooks and R scripts in old_stuff/, but these are either outdated or irrelevant for the paper. I've kept them here just in case...

Config files

There are two config files:

• config.yml - containing configuration (filepaths etc.) that you won't need to change
• userconf.yml - containing configuration that you might need to change, depending on e.g. where you have mounted the SEACO CH-20 RDSF