Nicolas D.

Data Scientist

830 dollar

Mon expérience

DataValue ConsultingSeptember 2019 - Présent

Voir plus

BelfiusMarch 2018 - Présent

Mission carried by Vertuo Conseil and Initio Belgium

Context: Belfius known as Dexia Bank Belgium up until 1 march 2012 is a company providing bank and insurance services. The company’s headquarter is located in Brussels. The main goal of the Customer Data Analytics office is to analyze and operate client’s data. Determine the different ways that Belfius customers are saving is one of the purpose of this office. Therefore, it is necessary to lead a data analysis to obtain a better business vision of savers firstly and secondly to propose a new segmentation that highlights the behavior associated to each way of saving.

Jupyther, Hive, Oracle
Python, Pyspark, SQL, HiveQL
Pandas, numpy, matplotlib, 3D, MLlib

-    Investment  product analysis: here we want to visualize the average number of different investment product  owned by the average client and how much money are invested on each of them (there are 6 investment product at Belfis: current account, savings account, securities account, investment insurance, pension savings and term account).
-    Investment product clustering.
-    Product transition analysis: Investigating on the typical order of product acquisition by calculating and visualizing the product transition matrix.
-    Invested amount evolution analysis: Customers categorization in 4 classes, increasing, decreasing, flat and no linear tendency.
-    Chanel analysis (mobile, web, tablet…)
-    2D and 3D data visualization.
-    Business intelligence: Synthesize all analysis in a note for the management.
-    Build the learning database for a clustering model on hive.
-    Clustering customers by using k-means and Hclust.

VERTUO ConseilMarch 2017 - Présent

Voir plus

ItelisAugust 2017 - January 2018

Mission carried by Vertuo Conseil.

Context: Itelis is a service company specialized in the health care area. Their main shareholders are AXA and Humanis. They offer to facilitate  access to medical prestation or equipment, while always guaranteeing quality and an optimized price. One of Itelis issues is to identify  fraudulent request from customers and temporarily block them in order to conscientiously analyze them and identify their true nature. The final objective is to deliver a decision algorithm based on machine learning in order to detect fraudulent request.

Langages/Tools/librairies :
Python, SQL
Jupyther, Oracle, Dataiku
Pandas, numpy, matplotlib, scikitlearn

-Data exploration and visualization
-Create and select features  that has the most effect on target.
-Correlation analysis between selected features.
-Run machine learning experiments and testing several algorithm such as random forest, xgboost, logistic regression, SVM, k-means and others.
-Implement custom machine learning code.
-Implement model performance and robustness criterion.
-Tuning the hyper-parameters of the estimators with GridSearch
-Cross validation
-Deploying machine learning solutions into production.
-Implement a fraud suspicion ladder.
-Wright a Data dictionary
Voir plus

PSA Banque France - CrediparMarch 2017 - May 2017

Mission carried by Vertuo Conseil

Context: The securitization team wants to launch the UK Capital Relief project. The joint venture between Santander UK and Banque PSA wants to decrease their RWA (Risk weighted asset ). The objective is to implement a synthetic securitization system. This project allows to transfer the risk to the shareholders without transferring any assets.

- Monitoring the English fund.
- Monitoring the Swiss fund.
- Reloading book validation.
- Swiss investor report validation.
- Framing the business needs while helping functional person in charge.
- Follow-up and update requirements
Voir plus

CRÉDIT LOGEMENTApril 2016 - September 2016


Context: Credit Logement Paris is a credit guarantee company. One of the risk they are dealing with is customers doing early repayment of their credit.  In order to first understand and then predict this phenomenon, Credit Logement Paris wants a statistical survey that highlight the conjectural and structural levers that influence early repayment and then create a model.


-Develop and implement databases by acquiring data from primary or secondary data sources around customers behavior and conjectural indication.
-Identify, analyze and interpret customers patterns doing early repayment.
-Present information using data visualization techniques.
-Work with management to prioritize business and information needs
-Build predictive models.

Mes compétences


Machine Learning, Matplotlib, NumPy, Pandas

Big Data

Spark, Hive, PySpark, Big Data



Machine Learning



Python, R Language, SQL


Data Science, Spanish, Management, Consulting, Data analysis, Analytics


Product transition analysis, French, English, Python Programming, Apache Hive

Application servers


Computer Tools

MS Excel, Microsoft Excel

Business Intelligence

Business Intelligence, SAS Base

Mes études et formations

Master 2, Mathématiques et Information - Université de Montpellier2015 - 2016

Master 2 Management des systèmes d'information et gestion du risque - IAE Montpellier2015 - 2016