Dexcom CGM Data Analyzer

Scripts that take a CSV export of Dexcom continuous blood glucose monitor (CGM) data and generate plots to visualize the data in a meaningful way.

Dexcom allows raw patient CGM data to be exported in CSV format. Follow the instructions given here to export data.

Add your CSV to the data/ directory and rename it export.csv. A sample CSV is provided, and will be used if no export.csv is provided.

Install dependencies:

pip3 install -r requirements.txt

Add your Dexcom CGM data CSV named export.csv to the data/ directory. Run a script from the main directory with:

python3 <script_name.py> [options]

dexcom_analysis.py

Generates histogram charts for long acting insulin, fast acting insulin, and carb times
Generates a scatter plot of average blood glucose values for each day
Displays charts in one window
-f, --filename to specify data filename
-s, --save to optionally save charts as PNG

histogram.py

Generates histograms displaying the frequency of fast-acting insulin, long-acting insulin, and carbohydrate times.

mean_glucose_plot.py

dexcom_analysis.py
- Saving the charts from the PyWebView on MacOS is not supported
- Same notes as the below two scripts
histogram.py
- This script groups data into 30 minute buckets
- This script excludes carb entries less than 5 grams. This is to exclude small "correction" food that is used to sightly blood glucose when low. The justification for this is to only track more significant meals and snacks. This is adjustable by changing this line in the script:
```
df = df.loc[df.carbs >= 5.0]
```
mean_glucose_plot.py
- Dexcom CGMs will record blood glucose levels below 45 mg/dL as 'Low'. This script converts those values to 40 for simplicity. Actual values could be lower. This is adjustable by changing this line:
```
df['glucose_value'] = df['glucose_value'].replace('Low', 40)
```

kjthorpe18/dexcom-data-scripts