In 2020, we created a survey to get community input on how we prioritize our projects. The results gave us a good sense of which items people were most interested in. The top three requests were:
- Model stacking! This ended up resulting in the award-winning stacks package by our intern Simon Couch.
- Model ops (deploying and monitoring models). We are currently working on this from multiple fronts.
- Adaptive resampling and better parallel processing. The finetune package solved the first part and some work on the tune package resolved the second.
Almost everything that respondents prioritized highly last year has either been completed or is currently in progress. Most of our time right now is spent on model ops, survival analysis, and case weights. That work will take us through the end of the year (at least). Since this survey seemed to work well last year, we’ve decided to use it again to prioritize features for next year.
Learn more at