Ecopetrol faced two significant challenges:

1. Underutilisation of Data: They needed to leverage vast amounts of data across various departments for prescriptive analysis. The task involved understanding the business with industry experts, identifying pain points, and proposing data-driven solutions.

2. Budget Forecasting Optimization: This required precise financial predictions for both CAPEX and OPEX. Multiple machine learning models, including regression and forecasting models, were developed and trained to provide these predictions.

Throughout the seven-month project, business analysts supported the understanding and continuous improvement of development efforts. Azure’s resources were utilised to build a robust architecture, integrating data from SAP HANA and SharePoint into a development environment using Logic Apps and Data Factory, and processing it in Azure Databricks. The project involved data scientists, AI designers, and business analysts, ultimately enhancing Ecopetrol’s financial planning and operational efficiency.