Datasetpublic

Pradipta/tagore_songs

📘 Tagore Song Corpus A dataset of Rabindranath Tagore’s songs with musical, chronological, and classificatory metadata 📝 Dataset Summary This dataset contains a curated corpus of Rabindranath Tagore’s songs (Rabindra Sangeet), including full Bengali lyrics and detailed metadata describing musical structure, period classification, and dates of composition. Each row corresponds to a single song. The dataset is drawn from a manually curated CSV of approximately… See the full description on the dataset page: https://huggingface.co/datasets/Pradipta/tagore_songs.

sourceHugging Faceapache-2.0updated 10mo agoView on Hugging Face

0likes11downloads

Dataset Card

📘 Tagore Song Corpus

A dataset of Rabindranath Tagore’s songs with musical, chronological, and classificatory metadata

📝 Dataset Summary

This dataset contains a curated corpus of Rabindranath Tagore’s songs (Rabindra Sangeet), including full Bengali lyrics and detailed metadata describing musical structure, period classification, and dates of composition.

Each row corresponds to a single song. The dataset is drawn from a manually curated CSV of approximately 2000 entries.

📂 Dataset Structure

Data Fields

Below are the exact columns present in the dataset:

Column Name	Description
`Unnamed: 0`	Row index from the original CSV (can be ignored).
`গানের কথা`	Song Lyrics — Full Bengali text of the song.
`পর্যায়`	Primary Classification — Tagore’s thematic/period classification (e.g., পূজা, প্রেম, প্রকৃতি).
`ভাগ`	Sub-classification — Finer category or thematic subdivision within `পর্যায়`.
`ক্রমিক সংখ্যা`	Serial Number — Catalog or sequence identifier associated with the song.
`বাংলা_সন`	Bengali Year of composition (if known), using Bengali numerals in the source.
`ইংরেজি_সন`	Gregorian Year of composition (if known), using Bengali numerals.
`রাগ`	Rāga (mode) — The melodic framework to which the song is set.
`তাল`	Tāla (rhythmic cycle) — The rhythmic pattern associated with the song.

📊 Dataset Statistics

—~2,000 songs
—All textual content is Unicode Bengali

🧭 Example Usage

python

from datasets import load_dataset

ds = load_dataset("Pradipta/tagore_songs")

print(ds["train"][0]["গানের কথা"])   # lyrics
print(ds["train"][0]["রাগ"], ds["train"][0]["তাল"])

📜 Source

The dataset is a structured corpus created from publicly available texts of Rabindranath Tagore’s songs along with manually curated metadata. All lyrics are in the public domain (Tagore passed away in 1941; copyright expired in 2001).

⚖️ Licensing

Tagore’s works are in the public domain. The dataset compilation is released under: Apache License, Version 2.0.

🔍 Notes on Data Quality

—Some songs may have incomplete metadata (e.g., missing rāga or date).
—Spelling/orthography follows the original sources; no normalization applied.

📬 Contact

Email: ppmitra@gmail.com