Skip to contents

Input data

Source: vignettes/linelisting_vs_aggregated.Rmd
linelisting_vs_aggregated.Rmd

Input data format

All *_model() functions in serosv require data argument as the input data to be fitted.

The package can handle both linelisting and aggregated data, and it infers the format from the structure of the input data frame. This means that input data is expected to follow a specific format.

For linelisting data: data must have 3 columns as follows

  • the age vector (column name specified via age_col parameter)

  • the vector of counts of sero positives of that age group (column name specified via pos_col parameter)

  • the vector is the total population of that age group (column name specified via tot_col parameter)

For aggregated data: data must have 2 columns as follows

  • the age vector of individuals

  • is the vector indicating the serostatus of that individual (column name specified via status_col parameter)

Example: Fitting linelisting and aggregated data using polynomial_model()

linelisting <- parvob19_fi_1997_1998[order(parvob19_fi_1997_1998$age), ]
aggregated <- hav_bg_1964

# View the 2 different data format
head(linelisting)
#>     age seropositive year gender parvouml
#> 1     1            0 2001      m        8
#> 2     1            0 2001      m        7
#> 3     1            0 2001      m        6
#> 406   1            0 2001      f       11
#> 407   1            0 2001      f        7
#> 408   1            0 2001      f        5
head(aggregated)
#>   age pos tot
#> 1   1   3  16
#> 2   2   3  15
#> 3   3   3  16
#> 4   4   4  13
#> 5   5   7  12
#> 6   6   4  15

# fit with aggregated data
model1 <- polynomial_model(aggregated, k=1)
plot(model1)

# fit with linelisting data
model2 <- polynomial_model(linelisting, k=1, status_col = "seropositive")
plot(model2)

Data transformation

serosv also offers function transform_data() to convert from linelisting to aggregated data. For more information, refer to Data transformation

transform_data(
  linelisting,
  stratum_col="age", 
  status_col="seropositive") %>% 
  polynomial_model(k=1) %>% 
  plot()