Decoding Emotions: EDA for Better Hiring

Introduction

This project evaluates the communication and emotional qualities of candidates through Exploratory Data Analysis (EDA). Using data from interview videos, we aim to extract actionable insights into emotional stability, communication effectiveness, and adaptability. The analysis helps identify key traits that are critical for recruitment decisions.

Key Analyses and Findings

Distribution of Emotions

• Overall Distribution: Neutral emotions dominate, covering nearly 50% of the spectrum, followed by fear, sadness, and happiness. Neutrality suggests emotional control and calmness, suitable for high-pressure roles(e.g. HFT roles,Data Analysis Roles etc..).

• Individual Distribution: Candidates with varied emotions (e.g., fear, sadness, surprise) indicate adaptability and creativity but may require further evaluation for leadership roles. (e.g. Product Manager ,Director etc)

Emotional Profiles of Candidates

• Candidates exhibiting a high percentage of neutral emotions demonstrate emotional stability, crucial for leadership or management roles.
• Positive emotions like happiness suggest enthusiasm and a positive outlook, ideal for collaborative roles such as sales or public relations.

Gaze and Blink Analysis

• Gaze Stability: Indicates focus and attentiveness. Candidates with steady gazes are better suited for detail-oriented roles.
• Blink Frequency: Excessive blinking may indicate stress, while too little suggests unnatural behavior. A moderate rate reflects calmness and engagement.

Heatmaps and Correlation Matrices

• Emotion Relationships:
  • Positive correlations: Anger and disgust show a slight positive relationship.
  • Negative correlations: Neutral emotions negatively correlate with fear and anger, highlighting the ability to handle stress.

• Emotion and Communication Overlap:
  • Confidence positively correlates with conciseness (0.72) and enthusiasm (0.62), indicating strong communication skills.

Visualizations

Key visualizations include:

• Bar charts illustrating dominant emotion distributions.
• Spider plots highlighting individual emotional strengths and weaknesses.
• Heatmaps showing relationships between emotions and communication traits.

Conclusion

From the emotional profiles analyzed, we found that candidates with dominant neutral emotions show strong emotional stability and resilience, making them suitable for high-pressure roles like leadership, business analysis, or data-driven jobs. Positive emotions, like happiness and enthusiasm, indicate adaptability and teamwork skills, which are ideal for creative and team-based roles like marketing or public relations.

Candidates with more varied emotions, such as fear or sadness, showed adaptability but might need further evaluation for roles requiring steady emotional control, like leadership or client-facing positions. There is also a clear link between confidence, conciseness, and enthusiasm, which highlights the importance of strong communication skills in leadership and high-pressure environments.

While emotions and transcripts alone don’t fully determine a candidate’s suitability, they provide useful insights into how well someone handles tough situations like an interview. This can help guide better decisions in matching candidates to the right roles.

---

Repository Structure

Emotion-Led-Candidate-Analysis/
├── README.md               # Overview and instructions
├── notebooks/
│   └── EDA.ipynb  # Jupyter Notebook
├── docs/
│   ├── EDA_Report.md       # Report in markdown
│   └── EDA_Report.pdf      # Original report in PDF
├── images/
│   └── visualizations.png  # Saved visualizations
│   └──Bar_graph_profile_diff_candids.png
│   └──bargraph.png
│   └──barplot.png
|    .....

Contact

For questions or contributions, feel free to reach out!