demo.R

# Copyright (c) Meta Platforms, Inc. and its affiliates.

# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.

#############################################################################################
####################         Meta MMM Open Source: Robyn 3.12.0       #######################
####################             Quick demo guide                     #######################
#############################################################################################

# Advanced marketing mix modeling using Meta Open Source project Robyn (Blueprint training)
# https://www.facebookblueprint.com/student/path/253121-marketing-mix-models?utm_source=demo

################################################################
#### Step 0: Setup environment

## Install, load, and check (latest) Robyn version, using one of these 2 sources:
## A) Install the latest stable version from CRAN:
# install.packages("Robyn")
## B) Install the latest dev version from GitHub:
# install.packages("remotes") # Install remotes first if you haven't already
# remotes::install_github("facebookexperimental/Robyn/R")
library(Robyn)

# Please, check if you have installed the latest version before running this demo. Update if not
# https://github.com/facebookexperimental/Robyn/blob/main/R/DESCRIPTION#L4
packageVersion("Robyn")
# Also, if you're using an older version than the latest dev version, please check older demo.R with
# https://github.com/facebookexperimental/Robyn/blob/vX.X.X/demo/demo.R

## Force multi-core use when running RStudio
Sys.setenv(R_FUTURE_FORK_ENABLE = "true")
options(future.fork.enable = TRUE)

## IMPORTANT: Must install and setup the python library "Nevergrad" once before using Robyn
## Guide: https://github.com/facebookexperimental/Robyn/blob/main/demo/install_nevergrad.R

################################################################
#### Step 1: Load data

## Check simulated dataset or load your own dataset
data("dt_simulated_weekly")
head(dt_simulated_weekly)

## Check holidays from Prophet
# 123 countries included. If your country is not included, please manually add it.
# Tipp: any events can be added into this table, school break, events etc.
data("dt_prophet_holidays")
head(dt_prophet_holidays)

# Directory where you want to export results to (will create new folders)
robyn_directory <- "~/Desktop"

################################################################
#### Step 2a: For first time user: Model specification in 4 steps

#### 2a-1: First, specify input variables

## For documentation of all arguments run ?robyn_inputs
InputCollect <- robyn_inputs(
  dt_input = dt_simulated_weekly,
  dt_holidays = dt_prophet_holidays,
  date_var = "DATE", # date format must be "2020-01-01"
  dep_var = "revenue", # currently one dependent variable allowed
  dep_var_type = "revenue", # "revenue" or "conversion" allowed
  prophet_vars = c("trend", "season", "holiday"), # "trend","season", "weekday" & "holiday" allowed
  prophet_country = "DE", # 123 countries included in dt_prophet_holidays
  context_vars = c("competitor_sales_B", "events"), # e.g. competitors, discount, unemployment etc
  paid_media_spends = c("tv_S", "ooh_S", "print_S", "facebook_S", "search_S"), # mandatory input
  paid_media_vars = c("tv_S", "ooh_S", "print_S", "facebook_I", "search_clicks_P"), # if provided,
  # Robyn will use this instead of paid_media_spends for modelling. Media exposure metrics
  # include typically, but not limited to impressions, GRP etc. If not applicable, use spend instead.
  organic_vars = c("newsletter"), # marketing activity without media spend
  factor_vars = c("events"), # indicate categorical varibales in context_vars or organic_vars
  window_start = "2016-01-01",
  window_end = "2018-12-31",
  adstock = "geometric" # geometric or weibull_pdf
)
print(InputCollect)

#### 2a-2: Second, define and add hyperparameters

## hyperparameter names are based on paid_media_vars. See right hyperparameter names:
hyper_names(adstock = InputCollect$adstock, all_media = InputCollect$all_media)

## Guide to setup & understand hyperparameters

## Robyn's hyperparameters have five components:
## - Adstock parameters (theta or shape/scale)
## - Saturation parameters (alpha/gamma)
## - Regularisation parameter (lambda). No need to specify manually
## - Time series validation parameter (train_size)
## - Beta coefficient penalty factor (_penalty). No need to specify manually

## 1. IMPORTANT: set plot = TRUE to create example plots for adstock & saturation
## hyperparameters and their influence in curve transformation.
plot_adstock(plot = FALSE)
plot_saturation(plot = FALSE)

## 2. Get correct hyperparameter names:
# All variables in paid_media_vars and organic_vars require hyperparameter and will be
# transformed by adstock & saturation.
# Run hyper_names(adstock = InputCollect$adstock, all_media = InputCollect$all_media)
# to get correct media hyperparameter names. All names in hyperparameters must equal
# names from hyper_names(), case sensitive. Run ?hyper_names to check function arguments.

## 3. Hyperparameter interpretation & recommendation:

## Geometric adstock: Theta is the only parameter and means fixed decay rate. Assuming TV
# spend on day 1 is 100€ and theta = 0.7, then day 2 has 100*0.7=70€ worth of effect
# carried-over from day 1, day 3 has 70*0.7=49€ from day 2 etc. Rule-of-thumb for common
# media genre: TV c(0.3, 0.8), OOH/Print/Radio c(0.1, 0.4), digital c(0, 0.3). Also,
# to convert weekly to daily we can transform the parameter to the power of (1/7),
# so to convert 30% daily to weekly is 0.3^(1/7) = 0.84.

## Weibull PDF adstock: The Probability Density Function of the Weibull also has two
# parameters, shape & scale, and also has flexible decay rate as Weibull CDF. The
# difference is that Weibull PDF offers lagged effect. When shape > 2, the curve peaks
# after x = 0 and has NULL slope at x = 0, enabling lagged effect and sharper increase and
# decrease of adstock, while the scale parameter indicates the limit of the relative
# position of the peak at x axis; when 1 < shape < 2, the curve peaks after x = 0 and has
# infinite positive slope at x = 0, enabling lagged effect and slower increase and decrease
# of adstock, while scale has the same effect as above; when shape = 1, the curve peaks at
# x = 0 and reduces to exponential decay, while scale controls the inflexion point; when
# 0 < shape < 1, the curve peaks at x = 0 and has increasing decay, while scale controls
# the inflexion point. When all possible shapes are relevant, we recommend c(0.0001, 10)
# as bounds for shape; when only strong lagged effect is of interest, we recommend
# c(2.0001, 10) as bound for shape. In all cases, we recommend conservative bound of
# c(0, 0.1) for scale. Due to the great flexibility of Weibull PDF, meaning more freedom
# in hyperparameter spaces for Nevergrad to explore, it also requires larger iterations
# to converge. When shape or scale is 0, adstock will be 0.

## Weibull CDF adstock (to be depreceated): The Cumulative Distribution Function of Weibull has two parameters,
# shape & scale, and has flexible decay rate, compared to Geometric adstock with fixed
# decay rate. The shape parameter controls the shape of the decay curve. Recommended
# bound is c(0, 2). The larger the shape, the more S-shape. The smaller, the more
# L-shape. Scale controls the inflexion point of the decay curve. We recommend very
# conservative bounce of c(0, 0.1), because scale increases the adstock half-life greatly.
# When shape or scale is 0, adstock will be 0.

## Hill function for saturation: Hill function is a two-parametric function in Robyn with
# alpha and gamma. Alpha controls the shape of the curve between exponential and s-shape.
# Recommended bound is c(0.5, 3). The larger the alpha, the more S-shape. The smaller, the
# more C-shape. Gamma controls the inflexion point. Recommended bounce is c(0.01, 1). The
# larger the gamma, the later the inflection point in the response curve.

## Regularization for ridge regression: Lambda is the penalty term for regularised regression.
# Lambda doesn't need manual definition from the users, because it is set to the range of
# c(0, 1) by default in hyperparameters and will be scaled to the proper altitude with
# lambda_max and lambda_min_ratio.

## Time series validation: When ts_validation = TRUE in robyn_run(), train_size defines the
# percentage of data used for training, validation and out-of-sample testing. For example,
# when train_size = 0.7, val_size and test_size will be 0.15 each. This hyperparameter is
# customizable with default range of c(0.5, 0.8) and must be between c(0.1, 1).

## 4. Set individual hyperparameter bounds. They either contain two values e.g. c(0, 0.5),
# or only one value, in which case you'd "fix" that hyperparameter.
# Run hyper_limits() to check maximum upper and lower bounds by range
hyper_limits()

# Example hyperparameters ranges for Geometric adstock
hyperparameters <- list(
  facebook_I_alphas = c(0.5, 3),
  facebook_I_gammas = c(0.3, 1),
  facebook_I_thetas = c(0, 0.3),
  print_S_alphas = c(0.5, 1),
  print_S_gammas = c(0.3, 1),
  print_S_thetas = c(0.1, 0.4),
  tv_S_alphas = c(0.5, 1),
  tv_S_gammas = c(0.3, 1),
  tv_S_thetas = c(0.3, 0.8),
  search_clicks_P_alphas = c(0.5, 3),
  search_clicks_P_gammas = c(0.3, 1),
  search_clicks_P_thetas = c(0, 0.3),
  ooh_S_alphas = c(0.5, 1),
  ooh_S_gammas = c(0.3, 1),
  ooh_S_thetas = c(0.1, 0.4),
  newsletter_alphas = c(0.5, 3),
  newsletter_gammas = c(0.3, 1),
  newsletter_thetas = c(0.1, 0.4),
  train_size = c(0.5, 0.8)
)

# hyperparameters <- set_default_hyppar("weibull_pdf", InputCollect$all_media)

# Example hyperparameters ranges for Weibull CDF adstock
# facebook_S_alphas = c(0.5, 3)
# facebook_S_gammas = c(0.3, 1)
# facebook_S_shapes = c(0, 2)
# facebook_S_scales = c(0, 0.1)

# Example hyperparameters ranges for Weibull PDF adstock
# facebook_S_alphas = c(0.5, 3)
# facebook_S_gammas = c(0.3, 1)
# facebook_S_shapes = c(0, 10)
# facebook_S_scales = c(0, 0.1)


#### 2a-3 (Optional) Saturation calibration - The curve calibrator - Beta feature

## Dummy input data for Meta spend. This is derived from Halo's reach & frequency data.
## Note that spend and response need to be cumulative metrics. Headers need to be
## kept the same as dummy dataset.

# data("df_curve_reach_freq")
#
# # Currently only supports curve_type = "saturation_reach"
# curve_out <- robyn_calibrate(
#   df_curve = df_curve_reach_freq,
#   curve_type = "saturation_reach"
# )
# curve_out$plot_reach_freq
#
# ## Reach & frequency calibration reasoning
# # When calibrating response saturation with reach and frequency, it's
# # recommended to use "reach 1+" for gamma lower bound and higher frequency bucket,
# # e.g. "reach 10+" in the dummy dataset, for gamma upper bound calibration.
# # Assuming an extreme situation when every user sees one impression and purchases
# # immediately, In such case, response curve equals to the reach 1+ curve. Gamma
# # controls the inflexion point of a Hill curve and a lower gamma means earlier
# # and faster saturation at a lower spend level. We believe that the cumulative
# # reach 1+ curve represents the earliest inflexion, thus we believe it's a
# # reasonable lower boundary for gamma for a selected channel. As frequency
# # increases, the inflexion point also delays and approaches the hidden true
# # response curve. In the dummy dataset, we've simulated reach 10+ to represent
# # upper bound for gamma. Use domain expertise to further narrow down or widen
# # the bounds. For alpha, we recommend to keep the value flexible as in default.
# facebook_I_gammas <- c(
#   curve_out[["curve_collect"]][["reach 1+"]][["hill"]][["gamma_best"]],
#   curve_out[["curve_collect"]][["reach 10+"]][["hill"]][["gamma_best"]])
# hyperparameters$facebook_I_gammas <- facebook_I_gammas

#### 2a-4: Third, add hyperparameters into robyn_inputs()
InputCollect <- robyn_inputs(InputCollect = InputCollect, hyperparameters = hyperparameters)
print(InputCollect)

#### 2a-5: (Optional) Effect size calibration

## Effect size calibration

# 1. Calibration channels need to be paid_media_vars or organic_vars names.
# 2. We strongly recommend to use Weibull PDF adstock for more degree of freedom when
# calibrating Robyn.
# 3. We strongly recommend to use experimental and causal results that are considered
# ground truth to calibrate MMM. Usual experiment types are identity-based (e.g. Facebook
# conversion lift) or geo-based (e.g. Facebook GeoLift). Due to the nature of treatment
# and control groups in an experiment, the result is considered immediate effect. It's
# rather impossible to hold off historical carryover effect in an experiment. Therefore,
# only calibrates the immediate and the future carryover effect. When calibrating with
# causal experiments, use calibration_scope = "immediate".
# 4. It's controversial to use attribution/MTA contribution to calibrate MMM. Attribution
# is considered biased towards lower-funnel channels and strongly impacted by signal
# quality. When calibrating with MTA, use calibration_scope = "immediate".
# 5. Every MMM is different. It's highly contextual if two MMMs are comparable or not.
# In case of using other MMM result to calibrate Robyn, use calibration_scope = "total".
# 6. Currently, Robyn only accepts point-estimate as calibration input. For example, if
# 10k$ spend is tested against a hold-out for channel A, then input the incremental
# return as point-estimate as the example below.
# 7. The point-estimate has to always match the spend in the variable. For example, if
# channel A usually has $100K weekly spend and the experimental holdout is 70%, input
# the point-estimate for the $30K, not the $70K.
# 8. If an experiment contains more than one media variable, input "channe_A+channel_B"
# to indicate combination of channels, case sensitive.

# calibration_input <- data.frame(
#   # channel name should be in paid_media_vars and / or organic_vars
#   channel = c("facebook_S",  "tv_S", "facebook_S+search_S", "newsletter"),
#   # liftStartDate must be within input data range
#   liftStartDate = as.Date(c("2018-05-01", "2018-01-01", "2018-07-01", "2017-12-01")),
#   # liftEndDate must be within input data range
#   liftEndDate = as.Date(c("2018-06-10", "2018-03-01", "2018-07-20", "2017-12-31")),
#   # Provided value must be tested on same campaign level in model and same metric as dep_var_type
#   liftAbs = c(40000, 120000, 60000, 200),
#   # Spend within experiment: should match within a 10% error your spend on date range for each channel from dt_input
#   spend = c(12000, 86000, 22000, 0),
#   # Confidence: if frequentist experiment, you may use 1 - pvalue
#   confidence = c(0.85, 0.8, 0.99, 0.95),
#   # KPI measured: must match your dep_var
#   metric = c("revenue", "revenue", "revenue", "revenue"),
#   # Either "immediate" or "total". For experimental inputs like Facebook Lift, "immediate" is recommended.
#   calibration_scope = c("immediate", "immediate", "immediate", "immediate")
# )
# InputCollect <- robyn_inputs(InputCollect = InputCollect, calibration_input = calibration_input)

################################################################
#### Step 2b: For known model specification, setup in one single step

## Specify hyperparameters as in 2a-2 and optionally calibration as in 2a-4 and provide them directly in robyn_inputs()

# InputCollect <- robyn_inputs(
#   dt_input = dt_simulated_weekly
#   ,dt_holidays = dt_prophet_holidays
#   ,date_var = "DATE"
#   ,dep_var = "revenue"
#   ,dep_var_type = "revenue"
#   ,prophet_vars = c("trend", "season", "holiday")
#   ,prophet_country = "DE"
#   ,context_vars = c("competitor_sales_B", "events")
#   ,paid_media_spends = c("tv_S", "ooh_S",	"print_S", "facebook_S", "search_S")
#   ,paid_media_vars = c("tv_S", "ooh_S", 	"print_S", "facebook_I", "search_clicks_P")
#   ,organic_vars = c("newsletter")
#   ,factor_vars = c("events")
#   ,window_start = "2016-11-23"
#   ,window_end = "2018-08-22"
#   ,adstock = "geometric"
#   ,hyperparameters = hyperparameters # as in 2a-2 above
#   ,calibration_input = calibration_input # as in 2a-4 above
# )

#### Check spend exposure fit and consider channel split if applicable
InputCollect$ExposureCollect$plot_spend_exposure

##### Manually save and import InputCollect as JSON file
# robyn_write(InputCollect, dir = "~/Desktop")
# InputCollect <- robyn_inputs(
#   dt_input = dt_simulated_weekly,
#   dt_holidays = dt_prophet_holidays,
#   json_file = "~/Desktop/RobynModel-inputs.json")

################################################################
#### Step 3: Build initial model

## Run all trials and iterations. Use ?robyn_run to check parameter definition
OutputModels <- robyn_run(
  InputCollect = InputCollect, # feed in all model specification
  cores = NULL, # NULL defaults to (max available - 1)
  iterations = 2000, # 2000 recommended for the dummy dataset with no calibration
  trials = 5, # 5 recommended for the dummy dataset
  ts_validation = FALSE, # 3-way-split time series for NRMSE validation.
  add_penalty_factor = FALSE # Experimental feature to add more flexibility
)
print(OutputModels)

## Check MOO (multi-objective optimization) convergence plots
# Read more about convergence rules: ?robyn_converge
OutputModels$convergence$moo_distrb_plot
OutputModels$convergence$moo_cloud_plot

## Check time-series validation plot (when ts_validation == TRUE)
## Read more and replicate results: ?ts_validation
# if (OutputModels$ts_validation) OutputModels$ts_validation_plot

## Calculate Pareto fronts, cluster and export results and plots. See ?robyn_outputs
OutputCollect <- robyn_outputs(
  InputCollect, OutputModels,
  pareto_fronts = "auto", # automatically pick how many pareto-fronts to fill min_candidates (100)
  # min_candidates = 100, # top pareto models for clustering. Default to 100
  # calibration_constraint = 0.1, # range c(0.01, 0.1) & default at 0.1
  csv_out = "pareto", # "pareto", "all", or NULL (for none)
  clusters = TRUE, # Set to TRUE to cluster similar models by ROAS. See ?robyn_clusters
  export = TRUE, # this will create files locally
  plot_folder = robyn_directory, # path for plots exports and files creation
  plot_pareto = TRUE # Set to FALSE to deactivate plotting and saving model one-pagers
)
print(OutputCollect)

## 4 csv files are exported into the folder for further usage. Check schema here:
## https://github.com/facebookexperimental/Robyn/blob/main/demo/schema.R
# pareto_hyperparameters.csv, hyperparameters per Pareto output model
# pareto_aggregated.csv, aggregated decomposition per independent variable of all Pareto output
# pareto_media_transform_matrix.csv, all media transformation vectors
# pareto_alldecomp_matrix.csv, all decomposition vectors of independent variables


################################################################
#### Step 4: Select and save the any model

## Compare all model one-pagers and select one that mostly reflects your business reality
print(OutputCollect)
select_model <- "5_257_2" # Pick one of the models from OutputCollect to proceed

#### Version >=3.7.1: JSON export and import (faster and lighter than RDS files)
ExportedModel <- robyn_write(InputCollect, OutputCollect, select_model, export = TRUE)
print(ExportedModel)

# To plot any model's one-pager:
myOnePager <- robyn_onepagers(InputCollect, OutputCollect, select_model, export = FALSE)

# To check each of the one-pager's plots
# myOnePager[[select_model]]$patches$plots[[1]]
# myOnePager[[select_model]]$patches$plots[[2]]
# myOnePager[[select_model]]$patches$plots[[3]] # ...

################################################################
#### Step 5: Get budget allocation based on the selected model above

## Budget allocation result requires further validation. Please use this recommendation with caution.
## Don't interpret budget allocation result if selected model above doesn't meet business expectation.

# Check media summary for selected model
print(ExportedModel)

# Run ?robyn_allocator to check parameter definition

# NOTE: The order of constraints should follow:
InputCollect$paid_media_selected

# Scenario "max_response": "What's the max. return given certain spend?"
# Example 1: max_response default setting: maximize response for latest month
AllocatorCollect1 <- robyn_allocator(
  InputCollect = InputCollect,
  OutputCollect = OutputCollect,
  select_model = select_model,
  # date_range = "all", # Default to "all"
  # total_budget = NULL, # When NULL, default is total spend in date_range
  channel_constr_low = 0.7,
  channel_constr_up = c(1.2, 1.5, 1.5, 1.5, 1.5),
  # channel_constr_multiplier = 3,
  scenario = "max_response"
)
# Print & plot allocator's output
print(AllocatorCollect1)
plot(AllocatorCollect1)

# Example 2: maximize response for latest 10 periods with given spend
AllocatorCollect2 <- robyn_allocator(
  InputCollect = InputCollect,
  OutputCollect = OutputCollect,
  select_model = select_model,
  date_range = "last_10", # Last 10 periods, same as c("2018-10-22", "2018-12-31")
  total_budget = 1500000, # Total budget for date_range period simulation
  channel_constr_low = c(0.8, 0.7, 0.7, 0.7, 0.7),
  channel_constr_up = 1.5,
  channel_constr_multiplier = 5, # Customise bound extension for wider insights
  scenario = "max_response"
)
print(AllocatorCollect2)
plot(AllocatorCollect2)

# Scenario "target_efficiency": "How much to spend to hit ROAS or CPA of x?"
# Example 3: Use default ROAS target for revenue or CPA target for conversion
# Check InputCollect$dep_var_type for revenue or conversion type
# Two default ROAS targets: 0.8x of initial ROAS as well as ROAS = 1
# Two default CPA targets: 1.2x and 2.4x of the initial CPA
AllocatorCollect3 <- robyn_allocator(
  InputCollect = InputCollect,
  OutputCollect = OutputCollect,
  select_model = select_model,
  #channel_constr_low = 0.1,
  #channel_constr_up = 10,
  # date_range = NULL, # Default: "all" available dates
  scenario = "target_efficiency",
  # target_value = 5 # Customize target ROAS or CPA value
)
print(AllocatorCollect3)
plot(AllocatorCollect3)

# Example 4: Customize target_value for ROAS or CPA (using json_file)
json_file = "~/Desktop/Robyn_202412181043_init/RobynModel-5_257_2.json"
AllocatorCollect4 <- robyn_allocator(
  json_file = json_file, # Using json file from robyn_write() for allocation
  dt_input = dt_simulated_weekly,
  dt_holidays = dt_prophet_holidays,
  # date_range = NULL,
  scenario = "target_efficiency",
  target_value = 2, # Customize target ROAS or CPA value
  plot_folder = "~/Desktop",
  plot_folder_sub = "my_subdir"
)

## A csv is exported into the folder for further usage. Check schema here:
## https://github.com/facebookexperimental/Robyn/blob/main/demo/schema.R

################################################################
#### Step 6: Model refresh based on selected model and saved results

## Must run robyn_write() (manually or automatically) to export any model first, before refreshing.
## The robyn_refresh() function is suitable for updating within "reasonable periods".
## Two situations are considered better to rebuild model:
## 1. most data is new. If initial model has 100 weeks and 80 weeks new data is added in refresh,
## it might be better to rebuild the model. Rule of thumb: 50% of data or less can be new.
## 2. new variables are added.

# Provide JSON file with your InputCollect and ExportedModel specifications
# It can be any model, initial or a refresh model
RobynRefresh <- robyn_refresh(
  json_file = json_file,
  dt_input = dt_simulated_weekly,
  dt_holidays = dt_prophet_holidays,
  refresh_steps = 4,
  refresh_iters = 2000,
  refresh_trials = 5
)
# Now refreshing a refreshed model, following the same approach
json_file_rf1 <- "~/Desktop/Robyn_202412181043_init/Robyn_202412181054_rf1/RobynModel-1_133_7.json"
RobynRefresh <- robyn_refresh(
  json_file = json_file_rf1,
  dt_input = dt_simulated_weekly,
  dt_holidays = dt_prophet_holidays,
  refresh_steps = 4,
  refresh_iters = 2000,
  refresh_trials = 5
)

# Continue with refreshed new InputCollect, OutputCollect, select_model values
InputCollectX <- RobynRefresh$listRefresh1$InputCollect
OutputCollectX <- RobynRefresh$listRefresh1$OutputCollect
select_modelX <- RobynRefresh$listRefresh1$OutputCollect$selectID

## Besides plots: there are 4 CSV outputs saved in the folder for further usage
# report_hyperparameters.csv, hyperparameters of all selected model for reporting
# report_aggregated.csv, aggregated decomposition per independent variable
# report_media_transform_matrix.csv, all media transformation vectors
# report_alldecomp_matrix.csv,all decomposition vectors of independent variables


################################################################
#### Step 7: get marginal returns

## Example of how to get marginal ROI of next 1000$ from the 80k spend level for search channel

# Run ?robyn_response to check parameter definition

## The robyn_response() function can now output response for both spends and exposures (imps,
## GRP, newsletter sendings etc.) as well as plotting individual saturation curves. New
## argument names "metric_name" and "metric_value" instead of "paid_media_var" and "spend"
## are now used to accommodate this change. Also the returned output is a list now and
## contains also the plot.

## Recreate original saturation curve
Response <- robyn_response(
  InputCollect = InputCollect,
  OutputCollect = OutputCollect,
  select_model = select_model,
  metric_name = "facebook_I"
)
Response$plot

## Or you can call a JSON file directly (a bit slower)
# Response <- robyn_response(
#   json_file = "your_json_path.json",
#   dt_input = dt_simulated_weekly,
#   dt_holidays = dt_prophet_holidays,
#   metric_name = "facebook_S"
# )

## Get the "next 100 dollar" marginal response on Spend1
Spend1 <- 80000
Response1 <- robyn_response(
  InputCollect = InputCollect,
  OutputCollect = OutputCollect,
  select_model = select_model,
  metric_name = "facebook_I",
  metric_value = Spend1, # total budget for date_range
  date_range = "last_10" # last two periods
)
Response1$plot

Spend2 <- Spend1 + 100
Response2 <- robyn_response(
  InputCollect = InputCollect,
  OutputCollect = OutputCollect,
  select_model = select_model,
  metric_name = "facebook_I",
  metric_value = Spend2,
  date_range = "last_10"
)
# ROAS for the 100$ from Spend1 level
(Response2$sim_mean_response - Response1$sim_mean_response) /
  (Response2$sim_mean_spend - Response1$sim_mean_spend)

## Example of getting organic media exposure response curves
sendings <- 30000
response_sending <- robyn_response(
  InputCollect = InputCollect,
  OutputCollect = OutputCollect,
  select_model = select_model,
  metric_name = "newsletter",
  metric_value = sendings,
  date_range = "last_10"
)
# Simulated cost per thousand sendings
response_sending$sim_mean_spend / response_sending$sim_mean_response * 1000

################################################################
#### Optional: recreate old models and replicate results

# From an exported JSON file (which is created automatically when exporting a model)
# we can re-create a previously trained model and outputs. Note: we need to provide
# the main dataset and the holidays dataset, which are NOT stored in the JSON file.
# These JSON files will be automatically created in most cases.

############ WRITE ############
# Manually create JSON file with inputs data only
robyn_write(InputCollect, dir = "~/Desktop")

# Manually create JSON file with inputs and specific model results
robyn_write(InputCollect, OutputCollect, select_model)

############ READ ############
# Recreate `InputCollect` and `OutputCollect` objects
# Pick any exported model (initial or refreshed)
json_file <- "~/Desktop/Robyn_202412181043_init/RobynModel-5_257_2.json"

# Optional: Manually read and check data stored in file
json_data <- robyn_read(json_file)
print(json_data)

# Re-create InputCollect
InputCollectX <- robyn_inputs(
  dt_input = dt_simulated_weekly,
  dt_holidays = dt_prophet_holidays,
  json_file = json_file)

# Re-create OutputCollect
OutputCollectX <- robyn_run(
  InputCollect = InputCollectX,
  json_file = json_file
  )

# Or re-create both by simply using robyn_recreate()
RobynRecreated <- robyn_recreate(
  json_file = json_file,
  dt_input = dt_simulated_weekly,
  dt_holidays = dt_prophet_holidays,
  quiet = FALSE)
InputCollectX <- RobynRecreated$InputCollect
OutputCollectX <- RobynRecreated$OutputCollect

# Re-export or rebuild a model and check summary
myModel <- robyn_write(InputCollectX, OutputCollectX, export = FALSE, dir = "~/Desktop")
print(myModel)

# Re-create one-pager
myModelPlot <- robyn_onepagers(InputCollectX, OutputCollectX, export = FALSE)
# myModelPlot[[1]]$patches$plots[[7]]

# Refresh any imported model
RobynRefresh <- robyn_refresh(
  json_file = json_file,
  dt_input = InputCollectX$dt_input,
  dt_holidays = InputCollectX$dt_holidays,
  refresh_steps = 6,
  refresh_mode = "manual",
  refresh_iters = 1000,
  refresh_trials = 1
)

# Recreate response curves
robyn_response(
  InputCollect = InputCollectX,
  OutputCollect = OutputCollectX,
  metric_name = "newsletter",
  metric_value = 50000
)