Analyzing Dataset of University Students’ Perceptions of AI Chatbots: Exploring Clustering and Insights into Learner Profiles
This project analyzes the dataset "Impact of Chatbots on University Students' Learning Experiences" from Kaggle (https://www.kaggle.com/datasets/jocelyndumlao/chatbots-impact-on-university-learning). The goal is to explore student perceptions of chatbots, perform sentiment analysis, cluster students based on usage patterns, and provide recommendations for chatbot improvements.
1. Data Processing and Visualization
Tasks: Preprocess and visualize demographic data, AI usage, and attitudes with pie charts, count plots, and bar charts.
Outcome: Cleaned and visualized dataset ready for further analysis.
2. Sentiment Analysis
Tasks: Analyze feedback sentiment (positive, negative, neutral) and extract keywords.
Outcome: Insights into student satisfaction and areas for improvement.
3. Frequent Users’ Perceptions
Tasks: Visualize the benefits and challenges perceived by frequent AI users (Q8, Q9).
Outcome: Bar charts showing frequent users' attitudes toward AI’s effectiveness and limitations.
4. Infrequent Users’ Concerns
Tasks: Visualize concerns and willingness of infrequent users toward AI (Q6, Q7).
Outcome: Insights into barriers to AI adoption among students.
5. Clustering Analysis
Tasks: Apply K-Means and hierarchical clustering to identify patterns in student attitudes.
Outcome: Clusters that highlight different user profiles based on AI perceptions.
Key Findings: Mixed perceptions of AI, with benefits for frequent users but concerns over over-reliance and ethics.
Advice: Improve AI functionality, address ethical concerns, target training for different user groups, and use continuous feedback for improvements.