File size: 1,217 Bytes
ede814c
c0d6d32
 
3707c1a
c0d6d32
ede814c
3707c1a
 
 
c0d6d32
 
 
 
3707c1a
 
 
c0d6d32
 
 
3707c1a
 
 
 
 
c0d6d32
 
 
 
 
 
 
 
 
 
f31b816
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
---
pretty_name: Pens & Pencils (Tabular, 6 columns)
tags: [tabular, education, augmentation]
license: cc-by-4.0
task_categories: [tabular-classification]
---

# Pens & Pencils (Tabular)

This dataset contains real-world measurements of pens and pencils collected manually,
with augmentation for experimentation in ML education.

## Features (6)
- **length_cm** (float)
- **weight_g** (float)
- **is_metallic** (0/1)
- **body_shape** (categorical; e.g., round, hexagonal, triangular)
- **has_clip** (0/1)
- **is_pen** (binary target; 1 = pen, 0 = pencil)

## Splits
- `original`: 30 manually measured rows (≥30)
- `augmented`: 300 synthetic rows (≥300), label-preserving jitter on numeric fields

## EDA
- See `assets/feature_dists.png` and `assets/target_dist.png` for summary distributions.

## Intended Use
Educational use for practice with tabular data, augmentation, and Hugging Face Datasets.

## Ethical Notes
All data collected is from personal, non-sensitive objects (pens/pencils). No PII.

## AI Usage
Some augmentation/plot code was assisted with GenAI, transparently documented in this notebook. The majority of code was referenced from the inclass tabular dataset example refactored for my use case