Grounded reality meets machine learning: A deep-narrative analysis framework for energy policy research.

Area of Science:

• Computational Social Sciences
• Energy Policy Research
• Narrative Analysis

Background:

• Text-based data, like narratives, are increasingly used in energy research and social science.
• Current policy applications often underutilize advanced text analysis tools.
• Manual narrative analysis faces challenges in scalability, repeatability, and bias.

Purpose of the Study:

• To illustrate deep-narrative analysis potential in energy policy research.
• To propose a nested methodology combining topic modeling and grounded theory.
• To address limitations of traditional narrative analysis in policy contexts.

Main Methods:

• Utilizing topic modeling from computational social sciences for narrative analysis.
• Applying a nested approach of topic modeling and grounded theory.
• Conducting a meta-analysis of bibliographic data on energy policy, narratives, and computational social science.
• Performing a proof-of-concept case study on energy externalities in Mumbai housing.

Main Results:

• The nested methodology offers advanced insight generation beyond frequentist approaches.
• It enables answering research questions based on text data structure.
• The approach systematically integrates qualitative evidence into policymaking.
• Demonstrates theoretical compatibility and practical application.

Conclusions:

• The proposed deep-narrative analysis methodology bridges the gap between qualitative evidence and policymaking.
• It offers a systematic, repeatable, and scalable approach to narrative analysis in energy policy.
• This multidisciplinary methodology enhances the exploitation of digital text resources for policy insights.