Tasks and Priorities

[ColtAllen]

This post will be actively maintained, so keep checking back and if there are any specific tasks or suggestions you want to work on for the Beta release and beyond, please let me know.

Rodrigorivera inspired me to create Projects for all outstanding tasks, which will be sorted by version release:

BETA RELEASE

Deprecate fitters module and fix known lifetimes bugs:

• Create Aesara Op for scipy.special.hyp2f1
• Investigate Monetary Value Calculations
• Add pareto_nbd_model.py module
• Add bg_bb_model.py module
• Add bg_nbd_covars_model.py module
• Raise pymc exception whenever model fails to converge
• Fix Plotting Dependencies on Calibration vs. Holdout Purchases
• Add matrix plotting method to BaseModel
• Type hinting for utils.py, plotting.py, and generate_data.py modules

RELEASE CANDIDATE

Polish up documentation and fix any bugs identified while working on beta release(s):

• Add arviz plotting suggestions and code to docs for model inference
• Add an explanatory chart to documentation summarizing model characteristics
• Create and add logo for library
• Add troubleshooting guide to documentation
• Resolve documentation theme
• Docstrings for predictive methods.
• CI/CD for auto-updates to project documentation
• Consider hosting docs on GitHub and emulating documentation style of ArViZ and bambi

FUTURE VERSION RELEASES

Implement additional BTYD Models, enhance plotting functions, documentation and library performance, etc.

• Add time-invariant covariates to ParetoNBDModel
• Add RFM segmentation utility
• Add Pareto/GGG Model
• Add PDO Model
• Add test(s) comparing predictions between hourly and daily aggregations.
• Add BG/NBD-k and MBG/NBD-k Models
• Add time-varying covariates to BetaGeoModel
• Add time-varying covariates to ParetoNBDModel? (Note this model could take days to train and necessitate checkpointing functionality to be added)
• Suggestions for any other models to add?
• Add tutorial notebook(s)
• Add sample_kwargs to BaseModel.fit
• Fix BetaGeo conditional probability calculation
• Review viability of modifying hyperprior RVs
• Enhance existing plotting functions with seaborn
• jax sampling via pymc to speed up model fitting times
• Comprehensive evaluation of model priors to determine best defaults
• Refactor utils.py
• Add code for mlflow model registration to documentation
• Refactor BetaGeoModel._conditional_probability_alive unit test to remove extraneous args.
• Refactor BetaGeoModel._conditional_expected_number_of_purchases_up_to_time unit test to remove extraneous args.
• Delete redundant load_cdnow_summary() function
• Try GammaGammaModel._log_likelihood variant

Projects

• Documentation Adventures (Discussion)
• Deprecate fitters Module (Discussion)
• Refactor utils.py (Discussion)
• Plotting Parea (Discussion)
• New Models to Add (Discussion)
• CI/CD (Discussion)
• Library Optimization (Discussion)

[clintdunn]

@ColtAllen Do you have more context on each of the tasks? I'd like to investigate a few but not sure what needs to be done

[ColtAllen]

@clintdunn I appreciate the early enthusiasm 👍 I've been writing descriptions for each of the above tasks. After the alpha release is published I'll post them all as issues for collaborators to review and work on.

I'm targeting this Fri, 17-Jun for the alpha release (and task/issue postings) because I'm going on vacation afterward and won't be able to resume work on lifetimes until July.

[clintdunn]

@ColtAllen sounds good!

[ColtAllen]

EDIT: I'm back from vacation now and will be working on all alpha release issues.

Just posted two issues if anyone wants to work on them 👍 I consider posterior parameter sampling for predictions to be the most important task, but it would be best if the issues I just posted were completed first as it will simplify the work required.

Also, if anyone wants me to plan a Zoom call next month to discuss ongoing and future work, let me know.

[ColtAllen]

I'm back from vacation and have reorganized work tasks by version release. I'll handle everything for the alpha release, but there are a lot of fairly straightforward tasks to work on for the beta I could use some help with, nearly all of which involve modules that will be left as-is while in alpha. More information about these tasks will be provided in the pertinent project discussions and posted as issues soon.

[ColtAllen]

First version of Beta release is now available for pip installation! I intend to publish new Beta releases as new models are added.

[zwelitunyiswa]

@ColtAllen This is great. I have been waiting for a bayesian CLV in python (a la BTYD Plus in R). This is fantastic. Count me in for contributing documentation and the like. I have no experience in Github, but this would be a great opportunity to learn.

[ColtAllen]

Thanks @zwelitunyiswa! The bambi contribution guide is a great overview on how to use GitHub for open source projects, and may inspire some future additions to this library's CONTRIBUTING.md file.

The documentation could benefit from some refinements before the official release is published, so any feedback you have on the layout, color scheme, ideas to add, etc. would be helpful. Currently the documentation uses the pydata_sphinx_theme, but I'm not opposed to returning to the sphinx_rtd_theme used in the old documentation. The Kedro documentation uses that same theme in fact, albeit heavily customized.

[ColtAllen]

Second version of Beta release is now available for pip installation! This one includes a Bayesian variant of the Gamma-Gamma model and a few bug fixes.

[ColtAllen]

Third version of Beta release is now available for pip installation! This one includes a Bayesian variant of the Modified BG/NBD model, a few bug fixes, and requested enhancements to the legacy fitters.py package.