Healthcare Facility Reviews with Sentiment Labels

70,597 user reviews of medical facilities, with 41,419 labeled as positive and 29,178 as negative. Each review includes a title, text content, and is categorized into one of 48 medical facility types.

Use Cases

• Train a text classification model on the content and title fields to predict the sentiment label.
• Analyze the distribution of positive and negative sentiment across the 48 category types of medical facilities.
• Perform topic modeling on the content field of negative reviews to identify common patient complaints.
• Build a multi-label classifier using the content field to predict both the sentiment and the category of a facility.

Strengths

• Contains 70,597 labeled reviews, providing a substantial text corpus.
• Includes explicit sentiment labels (positive/negative) for 100% of the 70,597 records.
• Categorizes reviews across 48 distinct types of medical facilities.

Limitations

• The dataset is monolingual (Russian), limiting applicability to other language contexts.
• No information is provided on the geographic or temporal coverage of the reviews.
• The source and method of data collection are not detailed, raising questions about potential sampling bias.

Provenance

Source

null

Collection Method

null

Time Range

null

Freshness

null

Geography

null