Skip to content

shahrzadkhayatbashi/ShahrzadKhayatbashi-20250113

Repository files navigation

Customer Service Insight Analysis

Overview

This repository contains an analysis performed as part of a case study task to uncover insights from customer service and order data for a travel agency. The goal was to identify patterns in customer service interactions and provide actionable recommendations to improve operational efficiency and customer satisfaction.

The analysis is structured into multiple Jupyter notebooks, each focusing on a specific step of the process, from data cleaning to advanced analysis. The findings are supplemented with visualizations and key insights to address the objectives outlined in the case study.


Repository Structure

├── notebooks/
│   ├── 01_clean_orders.ipynb             # Cleaning and preprocessing the Orders dataset.
│   ├── 02_clean_errands.ipynb            # Cleaning and preprocessing the Errands dataset.
│   ├── 03_extract_xes.ipynb              # Preparing data for process mining and event analysis.
│   └── 04_analysis.ipynb                 # Exploratory data analysis, visualizations, and insights.
├── Shahrzad-20250112-Presentation.pptx   # Presentation of the result
├── requirement.txt                       # Used Python Libraries. 
└── README.md                             # Project documentation.

Analysis Workflow

  1. Data Cleaning:

    • Checking data inconsistencies.
    • Cleaning data especially unifying revenueues into the same currency.
  2. Feature Engineering:

    • Extracted and transformed key features, including time intervals and categorical groupings.
  3. Exploratory Data Analysis:

    • Generated descriptive statistics and visualizations to identify trends and patterns.
  4. Insight Generation:

    • Focused on significant relationships identified by investigating correlations among different features.
    • Performed K-means clustering and process mining algorithms.

Visualizations


Actionable Recommendations

  • We may be able to optimize process for Partner A as they cause the same number of contact but with much less revenue comparing Partner C.
  • We may be able to further investigate how multiple systems used in making an order may affect the number of contacts. The data was not enough to identify actionable insight, so further investigation can be done with the help of more data.
  • We can further analyze customer conversation using NLP techniques to identify most common topics that customers asked in most problemestic category. If those information convey to customer in a good time, we may be able to reduce the contacts significantly.
  • With using NLP techniques for 19: Chat and 2: Mail, we can investigate the most repetitive patterns throught the chats and based on that implement chatbot, which customers can get their answer by bot.

How to Run

Please install a virtual environment based on the requriement file provided in the repository.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published