Chapter 1 · My EDA project

Author

Affiliation

Your Name

How to fill this chapter in This is your main portfolio piece. The numbered sections below follow the IMRAD structure of research: question ≈ introduction, getting the data + cleaning decisions ≈ methods, plotting the story ≈ results, and findings ≈ discussion. In the Reproducible Workflows course:

• Session 2 lab: find a dataset you care about on an open-data portal (data.gouv.fr or your own country’s), source it reproducibly by code, run an EDA, clean the data, and commit it to git.
• Session 3 work block: finish and polish the chapter, pin your data and environment, open a pull request for review, and publish the book to a URL.

Everything between the TODO markers below is a template to replace with your own content.

The question

Getting the data

The data is sourced by code from an open-data portal of your choice. Because the download lives inside the document, it is recorded in the analysis itself; anyone who re-renders this document gets the same data and the same results. Record the dataset’s publisher and URL in the text so the provenance is explicit.

The only portal-specific part is the fetch. This template shows two ways; use the one that fits your portal:

• {rdatagouv} (data.gouv.fr only, optional): search/download from R.
• Any portal by URL: download.file() works with any portal’s direct download link.

# Load what you use. {rdatagouv} only if you're on data.gouv.fr and want it.
# library(rdatagouv)    # search / download datasets from data.gouv.fr
library(janitor)        # clean column names, tabulate with tabyl()
library(naniar)         # visualise missing values
library(skimr)          # one-line data summary with skim()

# TODO: replace with the code you wrote in the Session-2 lab.
# Option A — data.gouv.fr via {rdatagouv}:
# ds <- get_dataset("my-dataset-slug"); url <- ds$resources$url[1]
# Option B — any portal by direct URL:
# url <- "https://www.example-portal.org/.../my-data.csv"
# download.file(url, "data/my-data.csv")

Optionally pin the fetched copy with {pins} so the chapter reads from a versioned local board instead of hitting the network on every render:

board <- pins::board_folder("pins/", versioned = TRUE)
pins::pin_write(board, dat, "my-data", type = "csv")
dat    <- pins::pin_read(board, "my-data")

A first look

# The skim() output gives you type, missingness and distributions at a glance.
# skim(dat)

# vis_miss(dat) — where is data missing, and why might that be?

Cleaning decisions

Plotting the story

Findings

---

This chapter is your semester’s centrepiece. Revisit it as you learn new techniques and make it your best work by the final gallery walk.