Prediction of Personality Type via Video Résumé and Job Recommendations Considering OCEAN, MBTI, and the Dark Triad Indicators

Project Overview

This project aims to predict an individual's personality type based on the OCEAN model using multiple domains. The solution is transparent and ensures high-speed operation — approximately 1 second per video (note that the model is quantized). Additionally, a list of professions has been compiled, mapped to personality types based on OCEAN, MBTI, and the Dark Triad.

Functional Features and Technical Solution

1. Multimodal Personality Analysis Model

   The model processes three types of data (text, audio, video), allowing for precise analysis of personality characteristics based on the OCEAN system and converting them into MBTI. This approach ensures:

   • Deep contextual understanding,
   • High noise resilience,
   • Enhanced emotional interpretation,
   • Decision transparency.

2. Transformation of OCEAN into MBTI and the "Dark Triad"

   Using empirical data and the distribution of OCEAN across each MBTI type, accurate conversion is implemented:

   • A list of professions was compiled and processed.
   • Professions were mapped to OCEAN and MBTI personality types.
   • For MBTI — threshold values correlated with OCEAN were used,
   • For the "Dark Triad" — coefficients for each OCEAN trait to build a profile for each characteristic.

3. Model Processing Speed

   High processing speed (~1 second per video) due to an optimized quantized model. Suitable for handling large datasets, critical for quick selection and ranking.

4. Two-Stage Job Vacancy Ranking

   We use BM25 search for textual vacancy analysis and personalize the ranking using OCEAN assessment. This allows finding the most relevant vacancies for each candidate.

---

Project Killer Features

• OCEAN-to-MBTI and Dark Triad Transformation: Scientifically grounded and accurate transformation enables expanded analysis options.
• Multimodal Approach: Analysis of video, audio, and text ensures maximum assessment accuracy.
• Two-Stage Job Ranking: Technology that enhances recommendation relevance.

---

Technology Stack

• Python, PyTorch, OpenCV, Librosa, Whisper, Streamlit
• SQLite, SQLAlchemy
• Transformers, FaceNet, MTCNN, Catboost, Sber LightAutoML

---

Solution Uniqueness

The project stands out for its precise integration of three modalities, the ability to translate the OCEAN system into MBTI and the Dark Triad, and a high level of interpretability for all participants (employers and candidates).

Graphical Interface

For Employees

• A personal dashboard for video uploads.
• Receiving textual feedback.
• Viewing job recommendations with a ranked vacancy list.
• Additional feature: earning on the platform.

For Employers

• Admin panel for uploading multiple videos.
• Specifying desired profession and personality type.
• Receiving a ranked list of candidates.

Conversion of OCEAN to MBTI and Other Systems

• Translation of results into MBTI types.
• Conversion to other personality assessment systems.
• Textual explanation of user strengths and weaknesses.

Project Structure

model/ — personality type prediction model. professions/ — mapping of professions. translation/ — tools for translating between OCEAN, MBTI, and other systems. app_with_model.py - application with model prediction (long initialization, memory requirements). app.py — main application.

Installation and Launch

Instructions for installing and launching the project will be provided in subsequent sections.

FFmpeg and ffmpeg-python

What is FFmpeg?

FFmpeg is a powerful tool for processing multimedia files. It enables numerous audio and video operations.

What is ffmpeg-python?

ffmpeg-python is a Python library providing a convenient interface to use FFmpeg's capabilities within Python code. It allows easy integration of media processing functions into applications.

Installing FFmpeg

macOS

1. Install Homebrew (if not already installed):

   /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)

2. Install FFmpeg:

   brew install ffmpeg

3. Check the installation:

   ffmpeg -version

Windows

1. Download FFmpeg: Go to the FFmpeg official website and select the Windows version. It is recommended to use the build from gyan.dev or another reliable source.

2. Unpack the archive: After downloading the archive (usually in the format .zip), unpack it.

3. Add FFmpeg to the PATH:

   • Press Win + R, type sysdm.cpl and press Enter.
   • Go to the Advanced tab and click on Environment Variables.
   • In the "System Variables" section, find the Path variable, select it and click "Edit".
   • Click "Create" and add the path to the bin folder inside the FFmpeg directory, for example: C:ffmpegbin.
   • Click "OK" to save the changes.

4. Check the installation: Open the command prompt (CMD) and type:

   ffmpeg -version

If the installation is successful, you will see information about the FFmpeg version.

Linux

To install FFmpeg on Linux, use your distribution's package manager.: bash sudo apt-get install ffmpeg

Installing ffmpeg-python library

After installing FFmpeg, you can install the ffmpeg-python library via pip: bash pip install ffmpeg-python

Done. Now you can use the app.

Note: FFmpeg is needed to run Streamlit, don't forget about it!

Launching a Docker container

Requirements

• Docker — Make sure that Docker is installed on your device. If Docker is not installed, download it from the official website: Docker.

Installation and Launch

1. Clone the repository:

   git clone https://github.com/nZiben/video_cv_matching.git
   cd video_cv_matching

2. Build a Docker image:

   docker build -t video_cv_matching_image .

3. Launch the Docker container:

   docker run -p 8501:8501 video_cv_matching_image

4. Open the app:

   • Go to the browser and open http://localhost:8501 to see the application interface.

Stopping and Restarting the container

Container stop

1. Open a new terminal.
2. Enter the command to find out the ID of the running container.:

   docker ps

3. Use the docker stop command to stop the container:

   docker stop <CONTAINER_ID>

   Replace the <CONTAINER_ID> with the ID of your container.

Restarting an existing container

To restart a previously stopped container:

1. Find his CONTAINER ID or name using:

   docker ps -a

2. Then use the command:

   docker start <CONTAINER_ID or name>

Restarting the container with reassembly

If you need to make changes to the image and reassemble the container:

1. Rebuild the image:

   docker build -t video_cv_matching_image .

2. Restart the container:

   docker run -p 8501:8501 video_cv_matching_image

Additional information

-Dependency Update: If you change dependencies, don't forget to update requirements.txt before reassembling the image.

• Using Docker Compose: If the project requires multiple services, you can add the docker-compose' file.yml for easier startup.