Robyn is an experimental, AI/ML-powered and open sourced Marketing Mix Modeling (MMM) package from Meta Marketing Science. Our mission is to democratise modeling knowledge, inspire the industry through innovation, reduce human bias in the modeling process & build a strong open source marketing science community.
After testing with few beta users on bl01 branch, it's good to land.
fix: several bugs when refreshing with robyn_refresh() - non-existing folders, cluster = FALSE, CSV exports, ts_validation, window_end inherited
fix: penalties hyperparameters checked correctly when recreating a model that used penalization
fix: similar plotting scales for refresh plots when performance is too large to improve readability (refresh_plots_json())
feat: new add_data parameter in robyn_write() to include raw data. When used, dt_input is not required to recreate model
feat: improved chain logic on refresh models. If user is not following chained directories but models are available in plot_folder, they will be loaded and used
feat: recreate InputCollect when RobynModel-models.json provided (no selected model)
This is still version 3.10.7 but will be CRAN's 3.11.0 if no bugs or issues are reported within a week or so after it lands (@gufengzhou)
After testing with few beta users on bl01 branch, it's good to land.
robyn_refresh()- non-existing folders, cluster = FALSE, CSV exports, ts_validation, window_end inheritedrefresh_plots_json())add_dataparameter inrobyn_write()to include raw data. When used,dt_inputis not required to recreate modelplot_folder, they will be loaded and usedThis is still version 3.10.7 but will be CRAN's 3.11.0 if no bugs or issues are reported within a week or so after it lands (@gufengzhou)