Ch.01 Mini Project — Attribute Auditor¶

Concept: Stevens' attribute taxonomy (Nominal → Ordinal → Interval → Ratio), discrete vs. continuous, appropriate summary statistics per scale.

What it does¶

Given any CSV file, attribute_audit.py will:

Infer each column's attribute type using name-hinting heuristics and value analysis
Classify discrete vs. continuous
Flag missing values per column
Print a column-type summary (bar chart of type distribution)

Usage¶

# Run the built-in demo (mirrors Ch.1 textbook dataset)
python attribute_audit.py

# Audit your own CSV
python attribute_audit.py my_dataset.csv

No external dependencies — pure Python 3 standard library only.

Sample Output¶

════════════════════════════════════════════════════════════════════════════════
  ATTRIBUTE AUDIT REPORT
  Source  : Demo (Ch.1 textbook dataset + extras)
  Rows    : 10   |   Columns: 7
════════════════════════════════════════════════════════════════════════════════

  Column  : Name
  Type    : Nominal       |  Discrete
  Notes   : 10 unique values.
············...

  Column  : Age
  Type    : Ratio         |  Discrete
  Notes   : Range [23.00, 82.00]
············...

  Column  : Temperature_C
  Type    : Interval      |  Continuous
  Notes   : Date/time/calendar-based column.
············...

  COLUMN TYPE SUMMARY
  Ratio      ███ 3
  Nominal    ██ 2
  Ordinal    █ 1
  Interval   █ 1

Key Concepts from Ch.1 Applied¶

Concept	Where it appears in code
Stevens' 4 attribute types	`infer_attribute_type()` return value
True-zero test for Ratio	`has_true_zero()` heuristic
Discrete vs. continuous	`is_integer_valued()` check
Missing value handling	counted and reported per column

Implementation Status¶

The script contains one complete infer_attribute_type() implementation and runs as a standalone demo or against a supplied CSV file.

Limitations & Future Ideas¶

Type inference is heuristic — column name + value analysis. Always verify manually for real datasets.
Extension: --force-type colname=Ordinal CLI flag to override inference per column.
Extension: compute_stats() — scale-appropriate statistics (mode for Nominal, median for Ordinal, mean/std for Interval/Ratio).
Extension: JSON output mode for use in downstream pipelines.