R data loader to generate a ZIP archive

The data loader below reads in the penguins data from a local file, performs multiple linear regressions, then outputs multiple files (with model estimates and predictions) as a ZIP archive.

# Attach required packages (must be installed)
library(readr)
library(tidyr)
library(dplyr)
library(broom)

# Data access, wrangling and analysis
penguins <- read_csv("src/data/penguins.csv") |>
    drop_na(body_mass_g, species, sex, flipper_length_mm, culmen_depth_mm)

penguins_mlr <- lm(body_mass_g ~ species + sex + flipper_length_mm + culmen_depth_mm, data = penguins)

mlr_est <- tidy(penguins_mlr)

mlr_fit <- penguins |>
    mutate(
        body_mass_g_predict = penguins_mlr$fitted.values,
        body_mass_g_residual = penguins_mlr$residuals
    )

# Write the data frames as CSVs to a temporary directory
setwd(tempdir())
write_csv(mlr_est, "estimates.csv")
write_csv(mlr_fit, "predictions.csv")

# Zip the contents of the temporary directory
system("zip - -r .")

The system function invokes the system command "zip - -r .", where:

• zip is the command for zipping files
• - means the archive is output to standard output (required for data loaders)
• -r, the recursive option, means all files are added to the zip archive
• . compresses the current working directory

Access individual files (estimates.csv, or predictions.csv) from the generated ZIP archive using FileAttachment:

const modelEstimates = FileAttachment("data/penguin-mlr/estimates.csv").csv({typed: true});

const modelPredictions = FileAttachment("data/penguin-mlr/predictions.csv").csv({typed: true});

We can quickly display the model estimates and predictions using Inputs.table:

Inputs.table(modelEstimates)

Inputs.table(modelPredictions)