Envisioning Insight-Driven Learning Based on Thick Data Analytics With Focus on Healthcare.

Area of Science:

• Health Informatics
• Social Media Analytics
• Qualitative Research Methods

Background:

• Healthcare institutions need to monitor social media for patient insights.
• Traditional big data analytics often fail to capture patient perspectives effectively.
• Simplistic textual analytics (e.g., bag-of-words) are insufficient for deep understanding.

Purpose of the Study:

• To explore advanced methods for detecting and analyzing patient insights from social media.
• To move beyond traditional analytics by employing qualitative and network analysis techniques.
• To infer actionable intelligence from social media conversations for healthcare improvement.

Main Methods:

• Utilizing "thick data analytics" to analyze social media conversations, focusing on relationships and communities within the data.
• Employing qualitative research methods alongside network analysis to identify patient pain points and needs.
• Applying machine learning and transfer learning to build predictive models from identified insights.

Main Results:

• Identification of 'conversation communities' within social media data to uncover hidden patient insights.
• Demonstration of techniques like conversation graph visualization and anomaly detection for deeper understanding.
• Development of a visionary approach combining qualitative insights with web analytics for actionable intelligence.

Conclusions:

• A novel approach combining "thick data analytics" and machine learning offers a more profound understanding of patient insights from social media.
• This method enables healthcare providers to better engage with patients and address their specific needs and pain points.
• The approach is being tested with geo-located Twitter data to assess the quality of care during the COVID-19 outbreak.