Julien Tchibinda

DATA SCIENCE PROJECT:

Analysis and prediction of vehicle CO2 emissions. Project aimed at identifying the technical characteristics of vehicles that most influence their CO2 emissions and developing a robust predictive model capable of estimating CO2 emissions for new vehicle types. Machine Learning project framework at DataScientest, focused on analyzing and predicting vehicle CO2 emissions, a crucial challenge for energy transition and environmental policy.

• Data Preparation and Cleaning: Rigorous data cleaning (handling missing values, detecting and treating outliers, correcting inconsistencies).
• Exploratory Data Analysis (EDA): Data exploration to identify trends and relationships. Use of visualization techniques (histograms, scatter plots, box plots) to identify the highest emitting vehicles.
• Predictive Modeling: Selection and implementation of several machine learning algorithms suitable for Regression, such as Linear Regression, Random Forests, and Gradient Boosting (XGBoost). Splitting data into training and testing sets to evaluate model generalization.
• Evaluation and Interpretation of Results: Evaluating model performance using key metrics such as RMSE (Root Mean Squared Error), MAE (Mean Absolute Error), and R-squared coefficient.
• Communication and Recommendations: Summarizing analysis and modeling results in a detailed report.

• Collection, analysis, cleaning of data from APIs.
• Internal. Data preprocessing and storage on Postgres DB.
• CI/CD management on operational environments.
• Application of data visualization best practices on incident management statistics.
• Structuring storytelling around incident phenomena (alerts, incidents, anomalies, trends).
• Optimized SQL querying and information structuring for decision support.
• Management of the developer team.
• Management of technical and functional breakdowns.
• Collaboration with non-technical profiles to translate needs into structured solutions.
• Participation in defining business KPIs in conjunction with operational management.
• Analysis of indicators on Datadog dashboards for business logs.
• Dashboard analysis with Power BI.
• Technical support for developers.
• Monitoring of good coding practices.
• Drafting technical documentation.

• Data analysis and implementation of VIVASTAR data ETL processes.
• Application development BUILD / RUN / BATCH.
• Development coaching / Pair Programming / Best practices.
• Management of deployments in QUAL/INT/PRE-PROD environments.
• Test automation (API, Unit, Integration, Business).
• Management of production incident tickets.
• Coaching business meetings.
• Version upgrade management.
• Deliverables & document management DAT / DAL / AUDIT.