AREZOO TIRDAD

Client: Joyride (white-label fleet management platform)

Time-series Forecasting – Optimization of vehicle distribution

• Developed ML models to predict demand, adapting to user behavior to ensure even bike availability versus stations and to further optimize the businesses operation.
• Applied SQL & Python, on a year’s worth of structured operational data, to analyze more than 320k records of completed rides among four different bike & scooter sharing companies in the UK, Europe, and Australia.
• Performed EDA pre- and post- model development to gather insight and explain performance variations of each fleet.
• Provided expert advice on data management to improve data modeling, reduce noise & enhance model performance.
• Created a Docker image hosting Flask scripts to serve the final model as an API and deployed it using AWS ECR and Fargate for testing and client demonstration.

Client: AweGoo (e-commerce retailer)

Data visualization and Reporting

• Transferred operational dashboards from Tableau to Microsoft Power BI, leading to more than 90% cost savings for the client while enhancing data visualization capabilities and providing real-time visibility and control.
• Developed Python scripts for ETL pipelines to transfer data from AWS RDS or Athena to Azure Blob Storage.
• Leveraged Azure Synapse data to build interactive dashboards for Profit Summary, Sales & Profit Rankings, Suppliers Orders Analysis, and Marketing Costs Analysis, presenting KPIs, trends and drivers, across multiple dimensions for operations in Canada & US.
• Automated the workflow between AWS and Azure, using AWS Lambda, triggered by new sales reports. Implemented scheduled weekly refresh, ensuring regular updates of the Power BI dashboards.