rcknplyr

Results 7 issues of rcknplyr

[model] Setup a Jupyter + Docker environment to run the model & visualizations without use of CLI

- Ensure data scientists and model authors can run the model & visualizations without the use of CLI - Identify interaction enablers/platform for Jupyter notebook based interactions - Define user...

[model] Identify libraries with better multi-processing performance and integrate into model

- Explore use of dask / joblib or other libraries - Identify parameters to decide which library to use - Integrate into model code

[model] Improve performance of Markov Chain Monte Carlo process

- Identify ways to improve performance of mcmc process without impacting the rest of the code - Evaluate parallel mcmc strategies

[ci/cd] Create E2E test with parameters/output approved by Penn and insert into release flow

- Create an E2E test with parameters/output approved by Penn - Include this test in ci/cd workflow to surface unexpected results from model changes

[app] Make package PIP installable

[model] Refactor model code: Simplify user experience for the authors of the model

- Remove classes that aggregate multiple parameters/functions. This makes code hard to follow for model authors - Take the larger functions and split them into smaller functions - Ensure a...

[model] Develop library for Bayesian fit SEIR model

- Preserve and carefully evolve the SEIR model created by Penn - Enable other organizations/applications to build different apps without impacting Penn - Limit refactoring of library to original authors...