Skip to content

Latest commit

 

History

History
123 lines (87 loc) · 4.22 KB

README.md

File metadata and controls

123 lines (87 loc) · 4.22 KB

Data Science Analysis (EDA) on Video Interview Dataset.

image

Welcome to the Interview Dataset Analysis repository! This project takes you on an in-depth journey through a dataset containing emotional, visual, and transcript data for 10 interview candidates. Using cutting-edge data analysis techniques, we explore each candidate's performance, helping you make data-driven hiring decisions!

📝 Dataset Breakdown

The dataset comprises three main components:

  • Emotion Data: Tracks emotional states like anger, happiness, and fear throughout the interview.
  • Gaze Data: Eye metrics like blink rate and gaze direction, measuring engagement and focus.
  • Transcript Data: Contains the transcription of each interview, including detailed metrics like speech speed and confidence levels.

Example Snippet of emotion_data.csv

movie_id,image_seq,participant_id,elapsed_time,distance,gaze,blink,eye_offset,angry,happy,sad
1,001,10,0.001,0.45,12,1,0.2,0,1,0

📊 Key Analysis Insights

Our analysis explores the following dimensions:

1. Gaze Analysis

Tracking candidates' eye contact provides valuable insights into their level of engagement and focus.

import plotly.express as px

# Plot gaze distribution for all candidates
fig = px.histogram(df, x="gaze", color="candidate_id", title="Gaze Distribution Across Candidates")
fig.show()

2. Emotion Analysis

We measured emotional responses like happiness and fear to gauge emotional intelligence (EQ).

# Generate emotion summary statistics
emotion_summary = df_emotion.groupby('candidate_id').agg(['mean', 'std'])

3. Speech & Confidence Metrics

Analyzing speech speed and confidence during the interview reveals how effectively candidates communicate.

# Calculate confidence score
df['confidence_score'] = (df['avg_confidence'] * 0.7) - (df['std_confidence'] * 0.3)

📈 Visualizing the Data

We created interactive plots using Plotly to visualize each candidate's performance across key metrics:

  • Gaze Sum Per Candidate: Total eye contact during the interview.
  • Emotion Variability: Tracking fluctuations in happiness, fear, and sadness.
  • Speech & Confidence Distribution: Insights into speaking style and clarity.
import plotly.graph_objects as go

# Plot confidence scores
fig = go.Figure(data=[go.Bar(x=candidates, y=confidence_scores)])
fig.update_layout(title="Confidence Score per Candidate", xaxis_title="Candidate", yaxis_title="Confidence Score")
fig.show()

🚀 Final Candidate Scores

After analyzing all the metrics, we scored each candidate across four categories:

  • Communication: Eye contact, speech speed, confidence, conciseness.
  • Emotional Intelligence: Stability and regulation of emotions.
  • Transcript Analysis: NLP techniques applied to transcriptions.
  • Final Hiring Decision: Based on the cumulative scores from all categories.

Candidate Score Example

# Final hiring decision example
final_scores = {
    'Candidate 1': 3.05,
    'Candidate 2': 4.23,  # Recommended for hire
    'Candidate 3': 3.23,
    ...
}

🔧 Tools & Techniques

This project leverages various powerful tools for data analysis and visualization:

  • Python (Pandas, Plotly, NLP Libraries): For data processing, statistical analysis, and visualizations.
  • VADER & NLP: Applied for sentiment analysis and keyword extraction.
  • Plotly: Interactive charts and data visualizations.
  • Google Colab: Interactive notebooks for running the analysis.

📚 Getting Started

Want to dive into the code? Here's how to get started:

  1. Clone this repo:

    git clone https://github.com/your-repo/EDA-interview-analysis.git
cd EDA-interview-analysis

  2. Install the required Python packages:

    pip install -r requirements.txt

  3. Run the analysis:

    jupyter notebook

  4. Open the notebook EDA_interview_analysis.ipynb and start exploring!

🎓 Contributing

We welcome contributions to improve this analysis! Feel free to open an issue or submit a pull request if you'd like to enhance the functionality or add new insights.