TiMMInsight

This repository provides appendix materials for the paper: "Towards Automated Multimodal Exploratory Data Analysis through Large Language Models".

Appendix of TiMMInsight

The appendix covers:

1. TIQC: LLM-based Interactive Query Clarification
   • Algorithm 4: LLM-driven query clarification
   • Algorithm 5: LLM-based ambiguity detection
   • Figure 15: Ambiguity detection prompt
2. Operators and Descriptions
   • Table 2: Logical and physical operators used in the system
3. Natural Language Query to Structured Sub-query
   • Example C.1 with prompts and JSON outputs (Figures 16-21)
4. Logical Operator Mapping
   • Prompt and JSON output (Figures 22-23)
5. Logical Plan Generation
   • Prompt and JSON output (Figures 24-25)
6. Physical Plan Generation
   • Cost model assumptions
7. User Study of TiMMInsight
   • Figure 27: Prompt used by ChatGPT-5 in the study

Note

For full technical context, read the main paper together with Appendix.pdf.