Generate Dutch Epidemiological Narratives

These functions use the active LLM to produce Dutch-language text based on data from the Medical Epidemiology department. Each call uses an isolated chat context so the main conversation history is not affected.

llm_describe() writes a Dutch epidemiological description of a dataset.
llm_interpret() interprets patterns and suggests epidemiological explanations.
llm_conclude() writes a Dutch conclusion paragraph from results or data.

llm_describe(
  data,
  topic = NULL,
  period = NULL,
  region = NULL,
  tone = c("formal", "accessible"),
  length = c("short", "medium", "long"),
  new = FALSE,
  ...
)

llm_interpret(data, hypothesis = NULL, comparison = NULL, new = FALSE, ...)

llm_conclude(..., new = FALSE)

Arguments

data: A data frame to describe or interpret. Summarised before sending to the LLM; raw rows are never sent verbatim.
topic: Topic or subject of the analysis, e.g. "E. coli resistance".
period: Time period covered, e.g. "2023-2024".
region: Geographic region, e.g. "Noord-Nederland" or "Groningen".
tone: Writing tone: "formal" (default) or "accessible".
length: Approximate length of the output: "short" (1 paragraph), "medium" (2-3 paragraphs), or "long" (4+ paragraphs).
new: Logical. If TRUE, re-initiates the LLM before the task call. Default is FALSE.
...: Additional arguments passed to initiate_llm().
hypothesis: Optional hypothesis to evaluate against the data.
comparison: Optional second data frame for comparison. Also summarised before sending.

Value

A character string with the generated Dutch text.