Hugging Face's logo

Datasets:
t22000t
/
climbing-resources-db

Tasks:
Other
Modalities:
Tabular
Text
Formats:
parquet
Languages:
English
Size:
< 1K
Tags:
climbing
rock-climbing
technique
training
books
youtube
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
climbing-resources-db / README.md
t22000t's picture
t22000t
fix: drop License bullet + cleaner Schema label
3778d25 verified
preview code
|
raw
history blame contribute delete
12.9 kB
metadata 
annotations_creators:
  - expert-generated
language:
  - en
license: cc-by-4.0
multilinguality:
  - monolingual
size_categories:
  - n<1K
source_datasets:
  - original
task_categories:
  - other
task_ids: []
pretty_name: Climbing Movement & Technique Resources
tags:
  - climbing
  - rock-climbing
  - technique
  - training
  - books
  - youtube
  - courses
  - podcasts
  - curated-dataset
configs:
  - config_name: default
    data_files: data/all_resources.parquet
dataset_info:
  features:
    - name: resource_id
      dtype: string
    - name: resource_type
      dtype: string
    - name: title
      dtype: string
    - name: creator
      dtype: string
    - name: year
      dtype: int64
    - name: disciplines
      dtype: string
    - name: technique_tags
      dtype: string
    - name: skill_level_min
      dtype: string
    - name: skill_level_max
      dtype: string
    - name: is_free
      dtype: bool
    - name: price_usd
      dtype: float32
    - name: url
      dtype: string
    - name: isbn
      dtype: string
    - name: language
      dtype: string
    - name: country
      dtype: string
    - name: description
      dtype: string
    - name: date_added
      dtype: string
    - name: last_verified
      dtype: string

Climbing Movement & Technique Resources

License: CC BY 4.0 Validate & Export URL Health

A curated, open-source database of 114 resources covering books, YouTube channels, online courses, websites/blogs, and podcasts, focused on rock climbing movement and technique. Every entry has been manually selected and verified, with book ISBNs cross-referenced against Amazon, AbeBooks, and publisher catalogues.

Covers all major disciplines: sport/lead, bouldering, trad, and top rope. Skill levels span complete beginner to elite. Every entry carries a 33-topic technique taxonomy so you can filter precisely for what you need.

Dataset Description

  • Repository: timothy22000/climbing-resources-db
  • Point of Contact: @timothy22000
  • Resource count: 114 (30 books, 25 YouTube channels, 20 courses, 20 websites, 19 podcasts)
  • Schema fields: 18, with controlled vocabularies for resource type, disciplines, and skill level
  • Technique tags: 33 across movement, hold shapes, grip variants, orientations, terrain, training, and mental skills

Why I built this

I wanted one place to search for climbing resources. Recommendations for specific techniques (heel hooks, drop knees, crack jamming) were scattered across YouTube comments, gym chat, and Reddit threads, with nothing I could actually filter or query. So I built it: a single tagged file I can grep to find every resource that matches what I'm working on right now.

At a Glance

Category Count Free % Free Notes
Books 30 n/a n/a Publication span 1993 to 2024; every ISBN-13 verified
YouTube channels 25 25 100% All verified active
Online courses 20 1 5% Altitude Climbing, TrainingBeta, Power Company, et al.
Websites & blogs 20 10 50% Mix of free articles and paid training programmes
Podcasts 19 17 89% Verified on Spotify
Total 114 53 46%

Schema

Each row in all_resources.csv / all_resources.parquet contains 18 fields:

Field Type Description
resource_id string Stable ID, e.g. BK-001, YT-012, CR-005
resource_type string book · youtube · course · website · podcast
title string Full title
creator string Author(s), channel name, or instructor
year int Year of publication or channel launch
disciplines string Pipe-separated: sport · bouldering · trad · top_rope · all
technique_tags string Pipe-separated tags from the 25-topic taxonomy
skill_level_min string beginner · intermediate · advanced · elite
skill_level_max string Same enum as skill_level_min
is_free bool Whether the resource is freely accessible
price_usd float Approximate price in USD (null if free or variable)
url string Primary URL
isbn string ISBN-13 (books only; all manually verified)
language string ISO 639-1 language code
country string Country of origin
description string 1–3 sentence editorial description
date_added string ISO 8601 date the entry was added
last_verified string ISO 8601 date the entry was last checked

Technique Tag Taxonomy (33 topics)

Movement positions & body mechanics: footwork · hip_rotation · drop_knee · flagging · heel_hooks · toe_hooks · knee_bars

Hold shapes: crimp_grip · slopers · pinches · underclings · jugs · pockets · edges

Grip variants (crimp sub-types): half_crimp · full_crimp · finger_drag

Hold orientations: gastons · side_pulls

Route / terrain types: slab · overhang · crack_climbing · offwidth

Movement quality: dynamic_movement · route_reading · strength

Technical gear skills: gear_placement · anchors

Training methods: finger_training · periodization

Mental & rehabilitation: mental_game · lead_falling · injury_rehab

Repository Layout 

climbing-resources-db/
├── README.md                     # This file + HuggingFace dataset card
├── LICENSE                       # CC BY 4.0
├── CONTRIBUTING.md               # How to add or fix entries
├── CHANGELOG.md                  # Version history
├── data/
│   ├── all_resources.csv         # Unified flat table (primary file)
│   ├── all_resources.parquet     # Snappy-compressed, typed (HF viewer-ready)
│   ├── books.csv
│   ├── youtube.csv
│   ├── courses.csv
│   ├── websites.csv
│   └── podcasts.csv
├── source/
│   └── climbing_resources_database.xlsx
├── scripts/
│   ├── export_csv.py             # XLSX → CSV + Parquet export
│   ├── validate.py               # CI validation script
│   └── check_urls.py             # Parallel URL health checker
└── .github/
    ├── workflows/
    │   ├── validate-pr.yml       # Schema validation on every PR
    │   ├── url-checker.yml       # Weekly URL health check
    │   └── export.yml            # Auto-regenerate CSVs on push to main
    └── ISSUE_TEMPLATE/
        ├── add_resource.md
        └── fix_data.md

Quick Start 

import pandas as pd

# CSV
df = pd.read_csv("data/all_resources.csv")

# Parquet (faster, typed)
df = pd.read_parquet("data/all_resources.parquet")

# All free resources tagged with footwork
free_footwork = df[
    (df["is_free"] == True) &
    (df["technique_tags"].str.contains("footwork"))
]

# Books for beginner–intermediate bouldering
bouldering_books = df[
    (df["resource_type"] == "book") &
    (df["disciplines"].str.contains("bouldering")) &
    (df["skill_level_min"] == "beginner")
]

# Everything covering crack technique, sorted by year
cracks = df[df["technique_tags"].str.contains("crack_technique")].sort_values("year")

Load from HuggingFace 

from datasets import load_dataset
ds = load_dataset("t22000t/climbing-resources-db")
df = ds["train"].to_pandas()

Highlighted Resources

Books: The Self-Coached Climber (Hague & Hunter, 2006), The Climbing Bible (Mobråten & Christophersen, 2020), 9 Out of 10 Climbers Make the Same Mistakes (Dave MacLeod, 2009), Smooth (Xian Goh, 2024), Crack Climbing: The Definitive Guide (Pete Whittaker, 2020), Performance Rock Climbing (Goddard & Neumann, 1993).

YouTube: Hooper's Beta, Lattice Training, Neil Gresham Masterclass, Wide Boyz, Movement for Climbers, Hannah Morris, Catalyst Climbing.

Courses: Dave MacLeod's Technique Course (Altitude Climbing), Fear of Falling (Hazel Findlay), Wide Boyz Global Crack School, Bouldering Blueprint (Magnus Midtbø).

Podcasts: The Nugget, TrainingBeta, Power Company, Careless Talk (Aidan Roberts), Climbing Gold (Alex Honnold), Circle Up! (Kyra Condie).

Source Data

The canonical source of this dataset is the GitHub repository at github.com/timothy22000/climbing-resources-db. All edits, additions, and corrections happen there via pull request, with schema validation and weekly URL health checks running automatically. The dataset on HuggingFace is published as a downstream mirror, refreshed alongside each tagged GitHub release.

Data flow: edits land on the GitHub main branch, the validate-pr.yml workflow checks schema conformance, and the data/all_resources.parquet file (regenerated from the master CSV) is then uploaded to this HuggingFace repo. The CSV under data/ on GitHub is always the source of truth for the Parquet here.

To contribute, file an issue, or report a broken link:

Data Quality & Verification

All 30 book ISBNs were manually verified against Amazon, AbeBooks, publisher websites, and library catalogues. During the initial verification pass: 13 ISBNs were corrected (the Climbing Bible series had systematic ISBN errors; several Falcon Guides had wrong edition numbers), and 6 entries were removed as unverifiable or fabricated. YouTube channels and podcasts were verified as active; online courses were confirmed available on their respective platforms. URL health checks run automatically on a weekly schedule.

Contributing

Contributions are welcome. Please read CONTRIBUTING.md before opening a pull request. Use the issue templates to suggest new resources or report data errors. Selection criteria: the resource must focus substantially on movement and technique, must be currently available, and books must have a verifiable ISBN.

Citation 

@dataset{climbing_resources_db_2026,
  title   = {Climbing Movement \& Technique Resources},
  year    = {2026},
  url     = {https://github.com/timothy22000/climbing-resources-db},
  license = {CC BY 4.0},
  note    = {Curated open dataset of 114 rock climbing technique resources.},
}

Licence

Creative Commons Attribution 4.0 International (CC BY 4.0). You are free to share and adapt the data for any purpose, provided you give appropriate credit and indicate any changes made.

Limitations & Bias

Coverage is weighted toward English-language resources from the US and UK. Non-English content (French, German, Spanish) is under-represented. YouTube and podcast links can break without notice; the weekly URL checker mitigates but does not eliminate stale entries. Resource selection involves editorial judgement and may under-represent creators with smaller audiences. Prices are approximate at the time of verification. Mobile apps (Crimpd, Mountain Project) and print magazines are out of scope for this version.