Skip to content

README.md

Latest commit

 

History

History
71 lines (47 loc) · 2.42 KB

README.md

File metadata and controls

71 lines (47 loc) · 2.42 KB
Raw

Monthly and Weekly Training Trends

Purpose

Demonstrate reproducible calendar aggregation from the reduced Run-All Activities CSV without exposing stable keys or personal data.

Intended question

How do activity frequency, total distance, and total duration differ by month and ISO week in the supplied sample?

Data and columns

  • File: input_sample.csv
  • Grain: one synthetic activity per row
  • Columns: activity_date_local, activity_type, sport_type, distance_m, duration_sec, avg_hr, activity_training_load
  • Deliberately omitted: garmin_activity_key, raw IDs, paths, hashes, exact real-world dates, coordinates, and memo text

Run method

Supply input_sample.csv and prompt.md to a trusted analysis tool, or reproduce the calculations with any CSV-capable local tool:

  • month = first seven characters of activity_date_local;
  • ISO week = the ISO calendar year and week for the date;
  • distance km = sum of distance_m divided by 1,000;
  • duration hours = sum of duration_sec divided by 3,600;
  • missing coverage = non-blank count divided by row count.

Prompt and result

The reusable request is in prompt.md. The reviewed expected analysis is in result.md.

Reading the output

Observed Facts contains calculations. Interpretation describes only the visible pattern. Uncertainty and Unsupported Conclusions prevent the sample from being treated as coaching, health, or causal evidence.

Facts

The sample contains eight rows, two represented months, complete distance and duration, and one missing value in each of two optional metrics.

Interpretation

The example can show a frequency-stable but volume-higher second month. It cannot explain the change.

What cannot be decided

Fitness, fatigue, training quality, data completeness, and an appropriate next training decision cannot be determined from these columns alone.

Privacy

All values and future dates are fictional. Do not replace the sample with real Run-All rows in a public copy. A real CSV must remain local, and any external derivative must remove garmin_activity_key and review date granularity.

Reproduction

Use the same input, formulas, grouping rules, and prompt sections. Calculated tables should match; generative wording may vary and is not byte-deterministic.

See the project-wide Analysis Handoff Specification.