Sliding window analysis, pt 3 - Githubissues

insilica / clinicaltrials-compliance-study

2 stars 0 forks source link

Sliding window analysis, pt 3 #25

Closed zmughal closed 2 months ago

zmughal commented 2 months ago

Add helper for running long-observe params
Use window period + cutoff as x values
Update to align long-observe more with Anderson 2015
Comment out sliding-window
Add more specific x labels with window dates
Update dvc.lock
Stratify windows by funding source
proportions() for computing percentage from count
Add note to env template
Extract all analysis to functions
Create helpers to normalize the anderson2015 data
Apply standardisation to other driver scripts
Fix plot output directory
Turn off logistic regression for now
Extract out processing of a single window
Leave window dates a character types
Optionally insert window names into the labels
Text adjustments for stacked bar plot

zmughal commented 2 months ago

fig result_reported_within stacked_area

Windows with 5 year cut-offs stratified by funding source.

zmughal commented 2 months ago

Original paper data:

fig result_reported_within stacked_area

Our pipeline:

fig result_reported_within stacked_area

zmughal commented 2 months ago

Comparison of studies on either side of rule effective date (binning up to 36+ months): fig result_reported_within stacked_area