explaingit

zkldi/lr2ir-dataset

★ 14Audience · developerComplexity · 2/5Setup · moderate

TLDR

A preserved dataset of the LR2 rhythm game leaderboard captured before its shutdown, containing 25 million score entries and 326,000 chart records in a compressed SQLite database.

Mindmap

mindmap
  root((lr2ir-dataset))
    What it contains
      25M score entries
      326K chart records
      Raw HTML pages
    Data fields
      Player scores
      Chart metadata
      Rankings and combos
    How to browse
      DB Browser for SQLite
      Decompress first
    Limitations
      No player profiles
      No replays
      No course data
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Things people build with this

USE CASE 1

Download and browse the LR2 leaderboard archive to look up historical scores and rankings for rhythm game charts.

USE CASE 2

Query the SQLite database to analyze player performance trends or chart difficulty statistics across 326,000 beatmaps.

USE CASE 3

Build a read-only website or search tool on top of the dataset so non-technical players can browse the preserved records.

USE CASE 4

Use the chart metadata to cross-reference LR2 scores with other rhythm game databases or difficulty rating systems.

Tech stack

SQLite

Getting it running

Difficulty · moderate Time to first run · 30min

The SQLite database is compressed to ~4 GB, decompression and DB Browser for SQLite are required to browse it.

No license information found in the repository.

In plain English

LR2IR was an online leaderboard and ranking service for the rhythm game LR2. When the site announced it would shut down permanently on May 31st 2026, this repository was created to preserve as much of the data as possible, captured two days before closure. The main dataset is a large SQLite database file, compressed to roughly 4 gigabytes. It contains over 25 million individual score entries from players around the world, along with data for more than 326,000 charts. A chart in this context is a single song or beatmap that players compete on. The database records each player's score, ranking, combo, and accuracy details, alongside metadata about every chart such as its title, artist, difficulty level, and play statistics. A second dataset contains raw HTML pages for every chart entry that appeared on the site. The README notes this is mostly redundant with the database, but is included for completeness. The data is not yet presented in a friendly way for non-programmers. To browse the main database, you need to decompress it and open it in a tool called DB Browser for SQLite, which provides a spreadsheet-like view. The repository authors plan to build a read-only website so anyone can browse the preserved records without needing technical knowledge. Not everything could be saved. Player profile pages, most replay files, and course data were not fully captured because a third party was flooding the site with traffic in its final days, making it unreachable for scraping.

Copy-paste prompts

Prompt 1
I have the lr2ir-dataset SQLite database open in DB Browser. Write me a SQL query that finds the top 10 highest-scoring players for a specific chart by BMS title.
Prompt 2
Using the lr2ir-dataset SQLite file, write a Python script that exports all scores for a given player ID into a CSV file with columns for chart title, score, rank, and accuracy.
Prompt 3
I want to build a simple read-only website to browse the lr2ir-dataset. Show me how to set up a FastAPI backend that queries the SQLite database and returns chart scores as JSON.
Prompt 4
Write a SQL query against lr2ir-dataset that finds the 20 most-played charts by total score entry count, along with each chart's artist and difficulty level.
Open on GitHub → Explain another repo

← zkldi on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.