Baseline data simulation

library(ggplot2)
library(data.table)
library(magrittr)
library(surveillance)
#> Loading required package: sp
#> Loading required package: xtable
#> This is surveillance 1.21.0; see 'package?surveillance' or
#> https://surveillance.R-Forge.R-project.org/ for an overview.
#> 
#> Attaching package: 'surveillance'
#> The following object is masked from 'package:data.table':
#> 
#>     year

Simulate 16 baselines

Simulation of 16 different time series of count data. Parameter values for data simulation are presented in Noufaily et al. (2019), table 1.

Each of the simulatation take into account a combination of differents properties of syndromic data such as:

Baseline frequencies of reports (determined by parameter alpha)
Linear trends (determined by parameterbeta)
Seasonal trends (determined by parameters gamma_1 and gamma_2),
Day-of-the week effects (determined by parameters gamma_3 and gamma_4)
Over dispersion (determined by parameter phi).

baseline_1  <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 1,
                                    weekly_pattern_n = 2,
                                    alpha = 6, 
                                    beta = 0, 
                                    gamma_1 = 0.2, 
                                    gamma_2 = 0.2, 
                                    gamma_3 = 0.5, 
                                    gamma_4 = 0.4, 
                                    phi = 2, 
                                    shift_1 = 29 
                                  )



baseline_1[,s:=1]

baseline_2  <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 1,
                                    weekly_pattern_n = 2,
                                    alpha = 0.5, 
                                    beta = 0, 
                                    gamma_1 = 1.5, 
                                    gamma_2 = 1.4, 
                                    gamma_3 = 0.5, 
                                    gamma_4 = 0.4, 
                                    phi = 1, 
                                    shift_1 = -167 
                                  )


baseline_2[,s:=2]


baseline_3  <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 0,
                                    weekly_pattern_n = 2,
                                    alpha = 5.5, 
                                    beta = 0, 
                                    gamma_1 = 0, 
                                    gamma_2 = 0, 
                                    gamma_3 = 0.3, 
                                    gamma_4 = 0.25, 
                                    phi = 1, 
                                    shift_1 = 1
                                  )

baseline_3[,s:=3]


baseline_4 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 0,
                                    weekly_pattern_n = 2,
                                    alpha = 0, 
                                    beta = 0, 
                                    gamma_1 = 0, 
                                    gamma_2 = 0, 
                                    gamma_3 = 0.3,
                                    gamma_4 = 0.25, 
                                    phi = 1, 
                                    shift_1 = 1
                                  )

baseline_4[,s:=4]

baseline_5 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 1,
                                    weekly_pattern_n = 2,
                                    alpha = 6, 
                                    beta = 0, 
                                    gamma_1 = 0.3, 
                                    gamma_2 = 2, 
                                    gamma_3 = 0.3, 
                                    gamma_4 = 0.5, 
                                    phi = 1.5, 
                                    shift_1 = -50 
                                  )

baseline_5[,s:=5]


baseline_6 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 1,
                                    weekly_pattern_n = 1,
                                    alpha = 1, 
                                    beta = 0, 
                                    gamma_1 = 0.1, 
                                    gamma_2 = 2, 
                                    gamma_3 = 0.05, 
                                    gamma_4 = 0.05, 
                                    phi = 1, 
                                    shift_1 = -50 
                                  )

baseline_6[,s:=6]


baseline_7 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 0,
                                    weekly_pattern_n = 1,
                                    alpha = 6, 
                                    beta = 0.0001, 
                                    gamma_1 = 0, 
                                    gamma_2 = 0, 
                                    gamma_3 = 0.6, 
                                    gamma_4 = 0.9, 
                                    phi = 1.5, 
                                    shift_1 = 0 
                                  )


baseline_7[,s:=7]

baseline_8 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 1,
                                    weekly_pattern_n = 1,
                                    alpha = 3, 
                                    beta = 0, 
                                    gamma_1 = 1.5,
                                    gamma_2 = 0.1, 
                                    gamma_3 = 0.2, 
                                    gamma_4 = 0.3, 
                                    phi = 1, 
                                    shift_1 = -150 )



baseline_8[,s:=8]

baseline_9 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 1,
                                    weekly_pattern_n = 1,
                                    alpha = 3, 
                                    beta = 0, 
                                    gamma_1 = 0.2, 
                                    gamma_2 = 0.1, 
                                    gamma_3 = 0.05, 
                                    gamma_4 = 0.15, 
                                    phi = 1, 
                                    shift_1 = -200 
                                  )

baseline_9[,s:=9]


baseline_10 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 1,
                                    weekly_pattern_n = 1,
                                    alpha = 5, 
                                    beta = 0, 
                                    gamma_1 = 0.2, 
                                    gamma_2 = 0.1, 
                                    gamma_3 = 0.05, 
                                    gamma_4 = 0.1, 
                                    phi = 1, 
                                    shift_1 = 0 
                                  )


baseline_10[,s:=10]


baseline_11 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 2,
                                    weekly_pattern_n = 1,
                                    alpha = 0.5, 
                                    beta = 0, 
                                    gamma_1 = 0.4, 
                                    gamma_2 = 0, 
                                    gamma_3 = 0.05, 
                                    gamma_4 = 0.15, 
                                    phi = 1, 
                                    shift_1 = 0 
                                  )

baseline_11[,s:=11]


baseline_12 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 2,
                                    weekly_pattern_n = 1,
                                    alpha = 9, 
                                    beta = 0, 
                                    gamma_1 = 0.5, 
                                    gamma_2 = 0.2, 
                                    gamma_3 = 0.2, 
                                    gamma_4 = 0.4, 
                                    phi = 4, 
                                    shift_1 = 57
                                  )


baseline_12[,s:=12]



baseline_13 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 2,
                                    weekly_pattern_n = 1,
                                    alpha = 2,
                                    beta = 0.0005, 
                                    gamma_1 = 0.8, 
                                    gamma_2 = 0.8, 
                                    gamma_3 = 0.8, 
                                    gamma_4 = 0.4, 
                                    phi = 4, 
                                    shift_1 = 57 
                                  )

baseline_13[,s:=13]



baseline_14 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 1,
                                    weekly_pattern_n = 4,
                                    alpha = 0.05, 
                                    beta = 0, 
                                    gamma_1 = 0.01, 
                                    gamma_2 = 0.01, 
                                    gamma_3 = 1.8, 
                                    gamma_4 = 0.1, 
                                    phi = 1, 
                                    shift_1 = -85 
                                  )

baseline_14[,s:=14]


baseline_15 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 1,
                                    weekly_pattern_n = 2,
                                    alpha = 3, 
                                    beta = 0, 
                                    gamma_1 = 0.8, 
                                    gamma_2 = 0.6, 
                                    gamma_3 = 0.8, 
                                    gamma_4 = 0.4, 
                                    phi = 4, 
                                    shift_1 = 29)

baseline_15[,s:=15]



baseline_16 <- csalert::simulate_baseline_data(
                                    start_date = as.Date("2012-01-01"),
                                    end_date = as.Date("2019-12-31"),
                                    seasonal_pattern_n = 0,
                                    weekly_pattern_n = 2,
                                    alpha = 6, 
                                    beta = 0, 
                                    gamma_1 = 0, 
                                    gamma_2 = 0, 
                                    gamma_3 = 0.8, 
                                    gamma_4 = 0.4, 
                                    phi = 4, 
                                    shift_1 = 1
                                  )
                                   

baseline_16[,s:=16]

baseline <- rbind(baseline_1,
                  baseline_2,
                  baseline_3,
                  baseline_4,
                  baseline_5,
                  baseline_6,
                  baseline_7,
                  baseline_8,
                  baseline_9,
                  baseline_10,
                  baseline_11,
                  baseline_12,
                  baseline_13,
                  baseline_14,
                  baseline_15,
                  baseline_16)

baseline[, label:=paste("Sim", s, sep=" ")]

q <- ggplot(baseline, aes(x = date, y = n))
q <- q + geom_line(lwd = 1)
q <- q + facet_wrap(~s, scales = "free")
q <- q + csstyle::scale_fill_cs(palette = "posneg")
q <- q + csstyle::theme_cs(legend_position = "bottom")
q

Weekly patterns

q <- ggplot(baseline, aes(wday, n, group=wday))
q <- q + facet_wrap(~s, scales = "free")
q <- q + geom_boxplot()
q

Monthly patterns

q <- ggplot(baseline, aes(wday, n, group=calmonth))
q <- q + facet_wrap(~s, scales = "free")
q <- q + geom_boxplot()
q

Beatriz Valcarcel

2022-04-20

Simulate 16 baselines

Each of the simulatation take into account a combination of differents properties of syndromic data such as:

Weekly patterns

Monthly patterns