Mission/open goal Description
Principal Data Analyst
Investigate a hidden product-quality issue by analyzing sales, support tickets, customer chat conversations, and product-related documents in the Promptathon database. Using GitHub Copilot Agent and SQL, the objective was to identify products with strong sales but unusually high support burden and low customer satisfaction, determine the root causes through evidence-based analysis, and provide actionable business recommendations.
Harness and model
GitHub Copilot Agent with GPT-5.5
Turn-by-turn journey
1. Prompt
Prompt: Read the Promptathon repository, explain the available missions, and recommend the best mission for a beginner.
Agent response/action: Reviewed the README and mission files, compared all available missions, and recommended the Principal Data Analyst mission.
Result: Selected the Data Analyst mission and understood the investigation workflow.
2. Prompt
Prompt: Guide me through the Data Analyst mission and inspect the available SQL database.
Agent response/action: Connected to the Promptathon SQL database using SQL MCP tools, listed available tables, and inspected the schema.
Result: Identified core tables including Products, SalesOrders, SalesOrderLines, SupportTickets, SupportChats, Docs, Customers, and Employees.
3. Prompt
Prompt: Identify the top-selling products based on revenue and quantity sold, then compare them with support tickets and customer satisfaction.
Agent response/action: Generated and executed SQL queries to aggregate sales data and correlate it with support ticket volume and satisfaction scores.
Result: Identified the Premium Short Sleeve Men's Top (SKU: ZCPTM-SS-M-BW) as the strongest product-quality risk candidate.
4. Prompt
Prompt: Investigate the root cause using support chats and customer documents.
Agent response/action: Analyzed support chat conversations and customer documents to identify recurring complaint themes.
Result: Found repeated complaints related to connectivity failures, wash durability, sensor issues, and customer returns.
5. Prompt
Prompt: Prepare a complete Promptathon submission with findings, SQL evidence, recommendations, architecture, and reflections.
Agent response/action: Generated a structured Markdown report summarizing the investigation, evidence, business recommendations, and lessons learned.
Result: Produced a complete GitHub Issue submission ready for review.
Completion
Bonus work
Beyond the core mission, I validated the AI-generated insights by reviewing SQL query outputs before drawing conclusions. I combined structured SQL analysis with qualitative evidence from support chats and customer documents to strengthen the investigation. I also documented the complete analytical workflow, business recommendations, and a Mermaid architecture diagram to produce a professional, evidence-based submission.
Mission/open goal Description
Principal Data Analyst
Investigate a hidden product-quality issue by analyzing sales, support tickets, customer chat conversations, and product-related documents in the Promptathon database. Using GitHub Copilot Agent and SQL, the objective was to identify products with strong sales but unusually high support burden and low customer satisfaction, determine the root causes through evidence-based analysis, and provide actionable business recommendations.
Harness and model
GitHub Copilot Agent with GPT-5.5
Turn-by-turn journey
1. Prompt
Prompt: Read the Promptathon repository, explain the available missions, and recommend the best mission for a beginner.
Agent response/action: Reviewed the README and mission files, compared all available missions, and recommended the Principal Data Analyst mission.
Result: Selected the Data Analyst mission and understood the investigation workflow.
2. Prompt
Prompt: Guide me through the Data Analyst mission and inspect the available SQL database.
Agent response/action: Connected to the Promptathon SQL database using SQL MCP tools, listed available tables, and inspected the schema.
Result: Identified core tables including Products, SalesOrders, SalesOrderLines, SupportTickets, SupportChats, Docs, Customers, and Employees.
3. Prompt
Prompt: Identify the top-selling products based on revenue and quantity sold, then compare them with support tickets and customer satisfaction.
Agent response/action: Generated and executed SQL queries to aggregate sales data and correlate it with support ticket volume and satisfaction scores.
Result: Identified the Premium Short Sleeve Men's Top (SKU: ZCPTM-SS-M-BW) as the strongest product-quality risk candidate.
4. Prompt
Prompt: Investigate the root cause using support chats and customer documents.
Agent response/action: Analyzed support chat conversations and customer documents to identify recurring complaint themes.
Result: Found repeated complaints related to connectivity failures, wash durability, sensor issues, and customer returns.
5. Prompt
Prompt: Prepare a complete Promptathon submission with findings, SQL evidence, recommendations, architecture, and reflections.
Agent response/action: Generated a structured Markdown report summarizing the investigation, evidence, business recommendations, and lessons learned.
Result: Produced a complete GitHub Issue submission ready for review.
Completion
Bonus work
Beyond the core mission, I validated the AI-generated insights by reviewing SQL query outputs before drawing conclusions. I combined structured SQL analysis with qualitative evidence from support chats and customer documents to strengthen the investigation. I also documented the complete analytical workflow, business recommendations, and a Mermaid architecture diagram to produce a professional, evidence-based submission.