Skip to content
Snippets Groups Projects
Select Git revision
  • main default protected
1 result
Compare History
Madjid Sadallah's avatar
Added AUTHORS file
Madjid Sadallah authored
53f39af6
Name Last commit Last update
db
elasticsearch
extensions
logstash
rs
setup
.env
.gitignore
AUTHORS.md
LICENSE.md
README.md
docker-compose.yml
README.md

MOBILES RecSys

The MOBILES RecSys is an autonomous recommendation system designed to enhance the user experience of international students using the MOBILES app. By providing personalized recommendations based on user behavior, interactions, and preferences, the system leverages advanced data analytics and geolocated data. MOBILES RecSys integrates seamlessly with the MOBILES app, which allows users to explore the city, collect geolocated data, and share annotations, enriching their urban experiences. This system uses behavioral indicators to generate personalized recommendations that are tailored to each user's unique interactions with the city.

Features

  • Personalized Recommendations: The system generates recommendations based on user interactions and behavioral indicators, tailoring suggestions for each user profile.

  • Geolocated Annotations: Users can create detailed, geolocated annotations, enhancing their urban exploration experience and sharing personal insights about their environment.

  • Behavioral Indicators: These indicators are calculated in four main stages: Onboarding and Integration, Active Usage Support, Quality Improvement, and Urban Discovery.

  • Containerized Architecture: MOBILES RecSys is deployed using Docker and Docker Compose, which ensures optimal resource management and simplifies deployment.

  • Data Storage and Processing:

    • MariaDB is used for user data storage.
    • kTBS records digital traces of user interactions.
    • Logstash processes user data and interaction logs, which are then indexed into Elasticsearch.

  • Dashboard: A user-friendly interface, developed in Ruby, allows users to configure and automate system actions such as data retrieval, indicator calculation, recommendation generation, and notification sending.

Installation

1. Modify Configuration Files

Update the .env file with your environment-specific variables:

ELASTIC_VERSION=8.11.3
ELASTIC_PASSWORD='mobilespassword'
LOGSTASH_INTERNAL_PASSWORD='mobilespassword'
MONITORING_INTERNAL_PASSWORD=''
BEATS_SYSTEM_PASSWORD=''
MYSQL_USER='mobiles'
MYSQL_PASSWORD='mobilespassword'
MYSQL_ROOT_PASSWORD='mobilespassword'
MYSQL_DB='mobiles'
KTBS_USER='mobiles'
KTBS_PASSWORD='mobilespassword'
KTBS_URL='https://KTBS_URL/tracesmobiles/@obsels'

2. Modify Configuration for Recommendation System

Update the rs/config/settings.py file to reflect the correct database and Elasticsearch configurations, ensuring that the system connects to the right services.

3. Start the Setup and Run Containers

First, run the setup to initialize the necessary services:

docker-compose up setup

Then, launch the application in detached mode:

docker-compose up -d

4. Access the Dashboard

Once the application is running, open your web browser and go to http://localhost:8080 to access the MOBILES RecSys dashboard. The dashboard provides a user interface to manage system settings, automate actions like data retrieval, indicator calculation, recommendation generation, and notification dispatch.

Project Structure

setup/:                        Contains setup scripts and service initialization configurations.
elasticsearch/:                Configuration for the Elasticsearch service.
logstash/:                     Configuration and pipelines for Logstash.
db/:                           Database initialization scripts.
rs/:                           Source code for the recommendation system and Flask-based web application.
config/:                       Configuration files for application settings.

Technical Details

Setup

  • Initialization: The setup/ directory includes scripts for initializing the necessary services (Elasticsearch, Logstash, MariaDB, kTBS). These scripts configure and set up the environment, ensuring that all required services are in place before starting the application.

Elasticsearch

  • Search and Indexing: Elasticsearch is used for storing, searching, and indexing user interaction data. It enables efficient retrieval of logs and other data necessary for generating accurate and personalized recommendations.

Logstash

  • Data Processing: Logstash plays a crucial role in the system's ETL (Extract, Transform, Load) pipeline. It processes interaction logs, user data, and other inputs, transforming them into a format suitable for storage in Elasticsearch. It extracts data from MariaDB and kTBS, processes it, and sends it to Elasticsearch for indexing.

MariaDB

  • Database Storage: All user data, including collected logs and interactions, is stored in a MariaDB database. MariaDB serves as the persistent storage solution for user information and application interactions, ensuring that data is available for analysis and recommendation generation.

kTBS

  • Trace Management: The kTBS instance records digital traces of user interactions within the app. These traces are crucial for understanding user behavior and usage patterns. Logstash processes the data from kTBS, transforming it into actionable insights that contribute to the recommendation system.

Recommendation System (rs)

  • Python and Flask: The recommendation system is implemented in Python and utilizes Flask to create a web application. This system calculates behavioral indicators based on user interactions and generates personalized recommendations based on these indicators.

  • Configuration: The rs/config/settings.py file contains all the necessary configurations for connecting to external services such as the database and Elasticsearch. It also defines settings for recommendation algorithms and data processing.

Dashboard

  • User Interface: The MOBILES RecSys dashboard is developed in Ruby and provides a rich, interactive interface for managing system configurations. Users can configure the behavior of the system, automate tasks such as data retrieval, manage recommendation generation, and send notifications.

Running the System

To run the system, follow these steps:

  1. Modify Configuration Files: Ensure that the .env file and rs/config/settings.py are correctly updated with environment variables and configurations.

  2. Run Setup: Initialize necessary services with the following command:

    docker-compose up setup

  3. Start the Application: After setup, start the application in detached mode:

    docker-compose up -d

  4. Access the Dashboard: Open your browser and navigate to http://localhost:8080 to access the MOBILES RecSys dashboard, where you can manage the system and monitor the recommendation processes.

Contribution

Contributions are welcome! Please reach out to us if you are interested in contributing to this project.

License

This project is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).

Acknowledgements

This project is supported by the ANR under grant number ANR-20-CE38-0009. We thank the members of the partner laboratories for their contributions to the project's conception and ideas.

Contact

  • Madjid Sadallah
  • Marie Lefevre

If you have any questions or need assistance, please contact the project members via their respective emails.