Skip to content

CLAUDE.md

Latest commit

 

History

History
45 lines (28 loc) · 2.03 KB

CLAUDE.md

File metadata and controls

45 lines (28 loc) · 2.03 KB

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

An R-based analysis pipeline for Apple Health step count data extracted from an iOS backup (2016–2025). The project is a completed capstone for the Google Data Analytics Certificate.

Running the Pipeline

Scripts must be run in order from the project root in an R session (or RStudio with the .Rproj open):

source("R/01-load-daily-steps.R")   # Extract from SQLite → data/steps_daily.csv
source("R/02-clean-daily-steps.R")  # Flag/remove anomalies → steps_daily_flagged.csv, steps_daily_clean.csv
source("R/03-plot-daily-steps.R")   # Generate charts → charts/*.png

To compile the final report:

rmarkdown::render("report/James_Sauls_SQL_R_Data_Analysis_Sample.Rmd")

External Dependency

Script 01 requires access to the raw SQLite database. The path is set via .Renviron (not committed):

HEALTHDB_PATH=/path/to/healthdb_secure.sqlite

Scripts 02 and 03 only need the CSVs in data/, so they can run without the database.

Architecture

Data flow: healthdb_secure.sqlite → 01 → steps_daily.csv → 02 → steps_daily_flagged.csv + steps_daily_clean.csv → 03 → charts/*.png

Key domain concept — gap-anchor days: HealthKit backfills step counts onto the last day before a multi-day recording gap, inflating those values. Script 02 detects these by computing gap_days between consecutive dates and flags/excludes any day where gap_days > 1. This is the central data quality decision in the project; the flagged CSV preserves them for transparency.

SQLite schema: Steps are data_type = 7 in the samples table, joined to quantity_samples on data_id. Timestamps use Apple's epoch (Unix + 978307200). Full schema notes are in notes/healthdb-schema-discovery.md.

Key Parameters (in 03-plot-daily-steps.R)

  • window_days = 90 — rolling average window (centered, 2-sided)
  • bin_width = 1000 — histogram bin width in steps