Meta-Labeler: Streamlined Conversation Labeling Tool

Meta-labeler is a powerful, user-friendly tool for processing and labeling conversational data for analytics and optimization. Designed for precision and ease of use, it integrates seamlessly into workflows, enabling users to extract insights and streamline lead management. With features like dynamic segmentation control, customizable metrics, and real-time progress tracking, Meta-labeler simplifies conversation analysis for actionable outcomes.

Table of Contents

• Features
• Prerequisites
• Getting Started
  • Setup
  • Input Files
  • Using LW Chrome Extension with Meta-Labeler
• Usage Instructions
  • Interface Overview
• Scoring Mechanisms
• Example Data
• Applications and Benefits
• Contributions
• License

Features

• Segmented Labeling: Classify conversations across sales funnel stages:
  • Intake: Initial interactions with leads.
  • Engaged: Active interactions and follow-ups.
  • Qualified: Leads ready to proceed further.
• Customizable Metrics:
  • Sentiment Score (1: Very Negative to 5: Very Positive).
  • Engagement Score (1: Minimal to 5: Very High).
  • Customer Effort Score (1: Very Low to 5: Very High).
  • Response Type (Manual, Templated, GPT).
• Color-Coded Segmentation: Differentiate stages visually with unique colors.
• Dynamic Segmentation Control: Choose labeling modes (e.g., Intake Only), and the tool greys out irrelevant segments.
• Data Export: Automatically save labeled data into a structured Excel file (labeled_conversations.xlsx).

Prerequisites

• Python Version: 3.8 or higher.
• Dependencies: Install the required libraries:

  pip install pandas openpyxl

• Conversation files: You'll need conversation files in JSON format with unique file IDs.

Getting Started

Extract and Import data

• Use a data extraction tool like LW-Chrome-Extension to get JSON files with the conversations to be labeled.

Setup

1. Clone the repository:

   git clone https://github.com/<your-username>/meta-labeler.git
   cd meta-labeler

2. Place your JSON files in a folder named conversations in the same directory.

3. Run the tool from a terminal (Shift + Right-click and choose PowerShell):

   

   python conversation-labeler.py

4. Input the file directory in the shell in order to start labelling.

Alternatively, you can also manually change the directory at:

json_path = Path(r"C:\Users\REPLACE-YOUR-FOLDER-HERE\conversations")

Input Files

Place JSON conversation files in the conversations folder. Example structure:

{
  "conversation_data": [
    {"timestamp": "DD MMM YYYY, 16:48", "sender": "Lead name", "message": "Hi", "is_automated": false},
    {"timestamp": "DD MMM YYYY, 13:32", "sender": "Responder name", "message": "Hi there Thanks for getting in touch", "is_automated": false}
  ],
  "parsed_at": "...",
  "total_messages": 2
}

Usage Instructions

Interface Overview

1. Select Segments to Label:

   

   • Use the radio buttons under Segmentation Control to select the labeling mode.
   • Non-selected segments are disabled and greyed out.

2. Label Data:

   

   • Assign metrics like Sentiment Score, Engagement Score, Customer Effort Score, and Response Type.

3. Export Data: Labeled data is saved automatically to labeled_conversations.xlsx.

Scoring Mechanisms

• Sentiment Score: Rates the tone of the lead's messages (1 = Very Negative, 5 = Very Positive).
• Engagement Score: Measures interaction depth (1 = Minimal, 5 = Very High).
• Customer Effort Score: Assesses interaction difficulty (1 = Very Low Effort, 5 = Very High Effort).
• Response Type: Classifies responses as Manual, Templated, or GPT-generated.

Example Data

The repository includes two sample JSON files in the example-conversations folder:

• conversation_1.json
• conversation_2.json

Using LW Chrome Extension with Meta-Labeler

Follow instructions in README to download the extension: https://github.com/PV-Bhat/LW-Chrome-Extension

Workflow

1. Start with LW Chrome Extension: Export conversations from Meta Business Suite into JSON format.
2. Import into Meta-Labeller: Use the exported JSON files as input for labeling.
3. Analyze and Refine: Label, analyze, and export the processed data to structured formats (e.g., Excel).

Leads Wizard Ecosystem

Meta-Labeller, along with LW Chrome Extension, forms the foundation of the Leads Wizard ecosystem, designed to bring efficiency and insights to lead and customer management workflows.

Applications and Benefits

Meta-labeler is ideal for:

• Customer Support Teams: Improving response quality and effort scores.
• Sales Teams: Streamlining lead qualification with data-driven insights.
• Data Analysts: Extracting actionable metrics from conversation data.

Contributions

Contributions are welcome! To contribute:

1. Fork the repository.
2. Create a feature branch (feature/your-feature-name).
3. Commit your changes.
4. Push the branch and submit a Pull Request.

License

This project is licensed under the MIT License. See LICENSE for details.