feat: Improve error handling in ExecutorAgent by enhancing prompt clarity and response guidance for insufficient data scenarios

README.md

@@ -168,12 +168,14 @@ User: "What about their locations?"
 AI: [Understands context from previous query]
 
 "Calculate Return Rate by Region. Don't give any chart"
-→ Returns only table (respects preference)
+→ ⚠️ Note: Sample dataset lacks return data - AI will explain limitation and suggest alternatives
 
 "hi" or "how are you"
 → Responds conversationally without triggering data analysis
 ```
 
+> **📝 Note on Test Prompt #10:** The challenge includes "Calculate Return Rate" but the Sample Superstore.csv doesn't contain return/refund data. The AI correctly identifies this limitation and explains what data would be needed. This demonstrates intelligent error handling.
+
 ---
 
 ## 🔑 Key Technical Features

backend/app/agents/executor.py

@@ -646,19 +646,20 @@ Do NOT import any system modules. Only use: pandas, matplotlib.pyplot, numpy.
 
             # Build the prompt for Gemini
             if is_insufficient_data:
-                prompt = f"""You are a data analysis assistant. The user asked a question that cannot be fully answered with the available data.
+                prompt = f"""You are a helpful data analysis assistant. The user asked a question that cannot be answered with the current dataset.
 
 User Question: {question}
 
-Available Data Schema:
-- Columns: {columns_info}
-- Shape: {data_shape}
+Available Data Columns: {columns_info}
 
-Analysis Plan (shows what's missing):
+Analysis Plan Notes:
 {plan}
 
-Please provide a helpful response explaining what data would be needed to answer this question, and if possible, suggest what analysis CAN be done with the available data.
-Keep your response concise (2-3 sentences)."""
+Provide a friendly response that:
+1. Explains what data/column is missing (e.g., "Returns" or "Return Status")
+2. Suggests what alternative analysis CAN be done with the current data (e.g., sales by region, profit trends, etc.)
+
+Keep response concise, helpful, and actionable (3-4 sentences max)."""
             else:
                 # Perform basic pandas analysis to get actual results
                 analysis_result = self._perform_pandas_analysis(df, question, plan)