L3Cube-MahaSent-MD

L3Cube-MahaSent-MD

MH Specific

Multi-domain Marathi sentiment analysis dataset with 60,000 samples across 4 domains, labeled for positive, negative, and neutral sentiment.

Build a Marathi product review analyzer for e-commerce
HomepageGitHub

Quick Start

import pandas as pd
df = pd.read_csv("MahaSent-MD.csv")
print(df.groupby("label").size())
# Output: negative 20000, neutral 20000, positive 20000
Modality
text
Size
60,000 samples
License
Format
CSV
Language
mr
Update Frequency
static
Organization
L3Cube, Pune
Category
language-nlp

Schema

FieldTypeDescription
sentencestringMarathi text sentence
labelstringSentiment label: positive, negative, or neutral
domainstringSource domain (e.g., news, social media, reviews, general)

Build With This

Marathi social media sentiment dashboard for brands operating in Maharashtra
Customer feedback classifier for Marathi-language support tickets
Political sentiment tracker for Marathi news and social media during elections

AI Use Cases

Sentiment analysisOpinion miningSocial media monitoringCustomer feedback classification
Last verified: 2026-03-07