counterfactual_mix: Counterfactual Mix

In monkeypostulate/MarketingAnalyticsR:

Description

For a given variable and scenario,it simulates a counterfactual mix, and it estimates the differeence between the actual mix and the counterfactual mix.

Usage

counterfactual_mix(
  data_input,
  col_names,
  group_names,
  baseline_period,
  counterfactual_period
)

Arguments

data_input	baseline and counterfactuals are provided in the data
group_names	The period when the counterfactual scenario is calculated.
baseline_period	The period where the baseline mix is assumed
counterfactual_period	the period where the counterfactuals are simulated
col_name	Model used to construct senarios

Value

Return Counterfactual mix

Examples

## Calculate counterfactual scenarios under different scenarios

data("sales_data")

sales_model1 <- lm(total ~
wday +
  month,
data = sales_data %>%
  filter(date < as.Date("2018-07-01"))
)

start_campaign <- as.Date("2018-07-01")
end_campaign <- as.Date("2018-07-25")
campaign_period <- start_campaign + seq(0, end_campaign - start_campaign)
baseline_period <- min(chosen_period) + seq(-14, 0, 1)


cf_scenario1 <- sales_data %>%
  mutate(adj_med_spend = min(adj_med_spend)) %>%
  estim_counterfactual(sales_model1, chosen_period)
cf <- data.frame("date" = chosen_period, "counterfactual" = cf_scenario1)

#
sales_data <- sales_data %>%
  left_join(cf, on = "date")


delta_shops <- sales_data %>%
  mutate(predictions = counterfactual) %>%
  counterfactual_mix(
    col_names = c("x", "y"),
    group_names = c("Shop x", "Shop y"),
    baseline_period = baseline_period,
    counterfactual_period = campaign_period
  )

delta_shops %>%
  filter(date %in% campaign_period) %>%
  group_by(group) %>%
  summarise(
    actual = sum(actual),
    counterfactual = sum(counterfactual),
    delta = sum(delta),
    delta_days = sum(delta) / length(campaign_period),
    delta_perc = 100 * sum(delta) / sum(counterfactual)
  )

monkeypostulate/MarketingAnalyticsR documentation built on Feb. 9, 2020, 12:15 a.m.