Kobla Legbedje

Data Scientist โ€ข Mathematics โ€ข Teaching
Transforming Data into Insights

About

Kobla Legbedje Profile

Kobla Legbedje

Data Scientist

Data Scientist with a strong background in mathematics, hands-on experience in data analysis and teaching, and expertise in Python, R, SQL, and data-visualization tools.

Passionate about innovation and solving complex problems through data-driven approaches. With 5 years of web-development experience using React.js, I combine technical expertise with analytical thinking.

Data Analysis

Statistical modelling and data visualisation

ML Engineering

Neural-network and deep-learning systems

Data Engineering

Data pipelines and database management

Programming

Proficient in Python, R and SQL

Expertise

Statistical Analysis
Predictive Modeling
Supervised/Unsupervised Machine Learning
Deep Learning
Natural Language Processing
Computer Vision
Time Series Analysis
A/B Testing
Data Cleaning
Feature Engineering
Agile/Scrum
DevOps
MLOps
API
EDA
PCA
Web Scraping
RAG

Certifications

โ˜๏ธ
Verified
Responsive Web Design
Responsive Web Design

Make webpages that respond to different screen sizes by building a photo gallery with Flexbox, and a magazine article layout with CSS Grid.

2022
advanced web application
UI Design
๐ŸŒ
Verified
British Airways's Data Science
British Airways's Data Science

Scraped and analysed customer review data to uncover findings. Built a predictive model to understand factors that influence buying behaviour.

2024
dREZ5sWsbSTkRe7RN
Airport Planning
Assumption Building
Communication
Data Modeling
Data Science
Data Visualisation
Machine Learning
PowerPoint
โ„๏ธ
Verified
Hands On Essentials Data Warehouse
Hands On Essentials Data Warehouse

Comprehensive training on Snowflake data warehouse essentials and best practices.

2025
136557028
Snowflake Databases
Snowflake Warehouses
Snowflake SQL Worksheets
Snowflake External Stage
SQL

Technology Stack

๐Ÿค–

Data Science & ML

โ–ผ

7 technologies

๐Ÿ
Python
๐Ÿ“Š
R
๐Ÿ”ฅ
PyTorch
๐Ÿผ
Pandas
๐Ÿ”ข
NumPy
๐Ÿงฎ
SciPy
๐Ÿค–
Scikit-learn
๐Ÿ“ˆ

Data Visualization

โ–ผ

5 technologies

๐Ÿ“ˆ
Plotly
๐Ÿ“Š
Tableau
โšก
Power BI
๐Ÿ“‰
Matplotlib
๐ŸŒŠ
Seaborn
๐Ÿ—„๏ธ

Databases

โ–ผ

4 technologies

๐Ÿ—„๏ธ
SQL
๐Ÿ˜
PostgreSQL
๐Ÿƒ
NoSQL
โ„๏ธ
Snowflake
๐ŸŒ

Web Development

โ–ผ

3 technologies

๐ŸŸจ
JavaScript
โš›๏ธ
React.js
โ˜•
Java
โ˜๏ธ

Cloud & DevOps

โ–ผ

3 technologies

๐ŸŒ
Azure
๐Ÿณ
Docker
๐Ÿ”ง
Git
๐Ÿ“‹

Project Management

โ–ผ

3 technologies

๐ŸŽฏ
Jira
๐Ÿ“
Confluence
๐Ÿ“—
Excel

Experience

Data Science & AI Intern
Somfy Group
Mar โ€“ Sep 2025

Analyzed connected-product data to propose new applications โ€ข anomaly detection โ€ข built HR chatbot.

Python
SQL
FastAPI
Data-Science Program
British Airways (Forage)
Oct 2024

Customer-review scraping, insight extraction, predictive model (+10 % sales potential).

Python
Mathematics Teacher
Institut de Micro-Informatique
2020 โ€“ 2022

Taught statistics & linear-algebra, designed exams, analysed results.

Statistics
Linear Algebra
Pedagogy
Independent Web-Developer & Consultant
2020 โ€“ 2025

Built modern, responsive React.js interfaces for several projects.

React.js
Next.js
Node.js

Projects

AirScope โ€“ Air Quality Intelligence
Interactive dashboard for monitoring and forecasting air pollution levels in real-time, personalized health advice and spatial visualization.

Technologies Used

React
Leaflet
Node.js
Express
Python
FastAPI
scikit-learn

Key Features

  • Real-time air pollution monitoring using OpenWeather API
  • Interactive map with pollutant details (PM2.5, NO2, O3, etc.)
  • Personalized health tips based on user profile (asthmatic, athlete, etc.)
50+
users
10K+
dataPoints
99.9%
uptime
Brain-Computer Interface Dashboard
Real-time visualization platform for neural network data analysis and monitoring with advanced analytics and machine learning insights.

Technologies Used

React
D3.js
WebSocket
Python
FastAPI
Redis

Key Features

  • Real-time neural network data visualization
  • Predictive analytics with ML models
  • Automated anomaly detection
  • Multi-user collaboration tools
25+
users
5K+
dataPoints
98.5%
uptime
E-commerce Churn Prediction with Amazon Scraping
Full-stack system that scrapes live Amazon data and predicts customer churn using machine learning and advanced data pipelines.

Technologies Used

Python
Selenium
BeautifulSoup
Pandas
Scikit-learn
PostgreSQL

Key Features

  • Real-time Amazon product scraping with rotating proxies
  • Data extraction: title, price, reviews, availability, competition
  • Feature engineering and real-time data processing
  • Churn prediction using machine learning models
  • Scalable architecture with anti-detection mechanisms
15+
users
50K+
dataPoints
97.2%
uptime
Universitรฉ de Lille AI Chatbot
Intelligent assistant for university students powered by RAG architecture, with a futuristic React interface and contextual knowledge base.

Technologies Used

React.js
FastAPI
LangChain
OpenAI
Chroma
PostgreSQL
JWT
Python 3.9+

Key Features

  • Real-time chat with RAG-powered responses
  • Context-aware answers from university documents and FAQs
  • Conversation history management
  • Student authentication with JWT-based sessions
  • Futuristic neon UI with animated transitions
100+
users
2K+
dataPoints
99.1%
uptime
TikTok Hashtag Analyzer โ€“ Pour Toi
Interactive dashboard for analyzing TikTok content performance under the 'Pour Toi' hashtag, with CSV import and rich visual insights.

Technologies Used

React
Recharts
Tailwind CSS
PapaParse
JavaScript (ES6+)
Lodash
Lucide React

Key Features

  • CSV file upload and parsing with PapaParse
  • Interactive charts for reach and engagement metrics
  • Automatic calculation of engagement rate
  • Responsive UI with Tailwind and custom icons
  • Real-time filtering and clean data visualizations
30+
users
1K+
dataPoints
99.5%
uptime
Air Quality Visualizer
Interactive web app to visualize air pollution data from Airparif via Data.gouv.fr, with dynamic CSV import and chart rendering.

Technologies Used

React
Chart.js
CSV Parser
Tailwind CSS

Key Features

  • 3-day PM2.5 prediction using machine learning
  • Interactive charts for pollutants (NOโ‚‚, PM10, Oโ‚ƒ...)
  • Clean and responsive UI
  • Support for multiple pollutant types
  • Open data integration from Data.gouv.fr / Airparif
20+
users
3K+
dataPoints
98.8%
uptime

Connect

Ready to build the future together? Letโ€™s connect.

Cluses, France
Built with v0