https://github.com/janduplessis883/friends-and-family-test-analysis

Interactive Streamlit app to make sense of FFT Feedback - for a responsive Primary Care Service
https://github.com/janduplessis883/friends-and-family-test-analysis

Last synced: 30 days ago
JSON representation

Interactive Streamlit app to make sense of FFT Feedback - for a responsive Primary Care Service

• Host: GitHub
• URL: https://github.com/janduplessis883/friends-and-family-test-analysis
• Owner: janduplessis883
• License: mit
• Created: 2023-10-28T21:58:16.000Z (about 1 year ago)
• Default Branch: master
• Last Pushed: 2024-07-01T02:08:12.000Z (5 months ago)
• Last Synced: 2024-07-05T14:31:30.256Z (4 months ago)
• Language: Jupyter Notebook
• Size: 20.1 MB
• Stars: 1
• Watchers: 1
• Forks: 1
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
![Image](https://github.com/janduplessis883/friends-and-family-test-analysis/blob/master/images/fftestabout.png?raw=true)

### Revolutionizing Patient Feedback Analysis in Healthcare

Welcome to **AI MedReview**, your powerful new dashboard for elevating healthcare providers' understanding and utilization of patient feedback. Our cutting-edge solution focuses on the essential Friends and Family Test (FFT), empowering you to extract deeper insights from this invaluable data source.

At the core of AI MedReview lies a transformative approach that goes beyond mere quantitative metrics. By leveraging advanced natural language processing and machine learning techniques, we unlock the nuanced sentiments behind patient responses. Our dashboard assigns detailed scores to each piece of feedback, painting a more comprehensive picture of patient satisfaction levels.

Through **sentiment analysis** powered by Hugging Face's `cardiffnlp/twitter-roberta-base-sentiment-latest model`, we precisely determine the emotional tone of patient comments, be it positive, negative, or neutral. This level of granular understanding enables you to celebrate areas of excellence and swiftly identify opportunities for improvement.

But we don't stop there. To protect patient privacy, we employ robust named **entity recognition (NER)** capabilities, utilizing the Hugging Face `dbmdz/bert-large-cased-finetuned-conll03-english model`. This ensures any personally identifiable information (PII) is seamlessly anonymized, safeguarding the confidentiality of your valuable data.

Furthermore, our innovative **zero-shot classification** approach, powered by the Facebook `BART-large-mnli` architecture, allows us to categorize patient feedback with remarkable accuracy – even without specialized healthcare training data. By carefully curating our classification labels, we achieved a striking 0.91 accuracy, demonstrating the remarkable versatility of this model.

This comprehensive suite of advanced analytics empowers healthcare providers like yourself to move beyond mere data presentation and unlock a clearer, more actionable understanding of patient experiences. Armed with these insights, you can drive continuous improvements, elevate service quality, and enhance patient outcomes.

Explore the AI MedReview dashboard today and experience the transformative power of data-driven decision-making in healthcare. Join us on our mission to revolutionize the way patient feedback is leveraged to deliver exceptional care.

Visit [AI MedReview on GitHub](https://github.com/janduplessis883/friends-and-family-test-analysis), where collaboration and contributions are warmly welcomed.

# [View the Streamlit App](https://friends-and-family-test-analysis-pqev4j3c9katnrlv8kktnb.streamlit.app/)