Create README.md

README.md

@@ -1,13 +1,65 @@
----
-title: Spss Formatting
-emoji: 🐠
-colorFrom: purple
-colorTo: yellow
+title: Feedback Topic Sentiment Transformer
+emoji: 📊
+colorFrom: blue
+colorTo: green
 sdk: gradio
-sdk_version: 5.34.2
+sdk_version: 4.19.2
 app_file: app.py
 pinned: false
-short_description: Turn a sandsiv+ export into SPSS format
----
+license: mit
+Feedback Topic & Sentiment Transformer
+This tool transforms feedback data with topic and sentiment columns into a binary matrix format where each unique topic becomes a separate column.
+Features
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+Multi-format Support: Accepts Excel (.xlsx, .xls), CSV, and tab-delimited text files
+Flexible Column Detection: Configurable prefixes for topic, sentiment, and category columns
+Binary Matrix Output: Transforms topics into columns with 0/1 values
+Sentiment Analysis: Associates sentiment scores (positive=1, neutral=0.5, negative=0) with each topic
+Data Analysis: Provides insights on topic frequency and sentiment distribution
+Download Options: Export results as Excel or CSV files
+
+How to Use
+
+Upload your data file containing feedback with topic and sentiment columns
+Configure column prefixes to match your data format:
+
+Topic columns (e.g., [**WORKSHOP] SwissLife Taxonomy)
+Sentiment columns (e.g., ABSA:)
+Category columns (e.g., Categories:)
+
+
+Click "Transform Data" to process your file
+Review the analysis (optional) to see topic frequencies and sentiment distributions
+Download the transformed file in your preferred format
+
+Input Format
+Your input file should have columns structured like:
+
+Topic columns: Contains the topic/category names
+Sentiment columns: Contains sentiment values (positive/negative/neutral)
+Text column: Original feedback text (optional)
+Recommendation column: Recommendation scores (optional)
+
+Output Format
+The transformed file will contain:
+
+feedback_id: Unique identifier for each feedback row
+Topic columns: Binary (0/1) indicating topic presence
+Sentiment columns: Numeric sentiment scores for each topic
+original_text: Preserved from input (if available)
+recommendation_score: Preserved from input (if available)
+
+Example Use Case
+Perfect for transforming customer feedback data with multiple topics and sentiments into a format suitable for:
+
+Machine learning models
+Statistical analysis
+Topic modeling
+Sentiment trend analysis
+
+Technical Details
+Built with:
+
+Gradio for the web interface
+Pandas for data processing
+Support for Excel and CSV formats