User-Engagement-Analysis

Introduction / Problem Statement

The Scholar Hub, a group of young European researchers, relies on WhatsApp as its principal communication channel for disseminating invitations and reminders regarding bi-weekly meetings held every Wednesday and Saturday at 8:00 pm. Due to a recent decrease in session attendance, Scholar Hub management is seeking the development of an interactive dashboard capable of offering swift insights into user engagement and providing recommendations to bolster weekly session attendance.

Goal of the Project

The goal is to analyze WhatsApp Group Chat history to unearth insights on user engagement, design a Power BI dashboard for the management to make quick decisions, and provide recommendations to improve user involvement.

Skills and Tools Used

• PowerBI
• DAX
• Project Management: Trello Board
• Data Visualization
• Data Cleaning
• Data Manipulation and Exploration

Task Overview

Here is a concise high-level review of the critical tasks to achieve the project goal

• Wrote a Python script that handles and transforms chat history data, initially stored as a TXT file, into a structured CSV format.
• Loaded the CSV file into Power BI query for further data manipulation.
• Used DAX to create custom calculations and aggregations such as word volumes, chat volumes, most prevalent words, etc.
• Designed an interactive dashboard, enabling managers to make data-driven decisions on user engagement.

Outcome

The dashboard shows different chats that pleasantly communicate with the decision-makers, as depicted below. It has various options to filter down the information for a granular view and a clearer insight into the group dynamics.

Insights

Key findings from the analysis include:

• Approximately 62% of the chat and 65% of the words are generated by a core group of 10 members.
• The most frequently used words include “session,” “great,” “tonight,” “good,” “everyone,” and “join,” showing most of the chat is about reminders or feedback about the session.
• Sender ID 12 is the most active member, contributing 7% of the chats and typing 17% of the words.
• A significant portion (46% of the chat and 49% of the words) occurs between 8 pm and 10 pm, with a peak at 9 pm corresponding to the meeting period.
• The top 3 sender IDs, responsible for 28% of the chat volume, have not been active in the last 2 weeks. The highest-ranked sender was last seen 22 days ago, and the second-highest was last seen 52 days ago.
• Maximum engagement was noted in July, accounting for 17% of the chat volume, while February recorded the lowest engagement at 3%.
• There has been a consistent decrease in chat volume since July, corresponding to a decline in activity among many top contributors.

Recommendations

Based on the findings, here are some recommendations to improve user engagement:

• While a core group of 10 members is highly active, encourage others to participate actively by promoting inclusive and engaging discussions.
• Recognize and incentivize the top contributors, especially Sender Id 12, to maintain their high level of engagement. This could involve acknowledging their contributions, providing special privileges, or creating a reward system.
• Diversifying content can attract different users with varied interests, increasing overall engagement.
• Reach out to the top 3 sender IDs not seen in the last 2 weeks. Encourage them to re-engage by highlighting recent interesting discussions.
• Plan special events or initiatives, such as themed discussions, Q&A sessions, or expert panels, to create excitement and boost engagement.
• Given the consistent decrease in chat volume since July, investigate the reasons behind the decline and take proactive measures, such as introducing new features, addressing user concerns, or refreshing the platform.
• Establish a constructive feedback mechanism to understand user preferences and areas for improvement.

Ready to unlock the full potential of your data? I’m excited to collaborate and help your business thrive through data-driven insights. Get in touch!