Data Science team enablement for a large global fortune 500 organization
Most organizations have started to view data science and data engineering as a strategic business function and are developing strong internal capabilities. For developing data science capability, it is critical to have the right people (and team structure), the right process and build the right tools. Actify Data Labs enables organizations to build data engineering and data science capability by helping train people, build processes and provide tools & utilities.
About the Client
Petronas (Petroliam Nasional Berhad), a Fortune 500 company, is one of the largest oil and gas company in the world. As part of its digital strategy Petronas is creating a comprehensive data management and data science organization. The data science team at Petronas envisages to leverage the power of data to unlock new insights and optimize performance across the value chain. The company is developing deep data science expertise mostly using open source technology.
The enablement initiative received 100% endorsement and a very positive feedback from each participant and from the data science leadership at Petronas. The utilities provided by Actify Data Labs helped reduce the time required for completing a data science project (particularly in the area of IOT analytics) by as much as 20%.
Petronas is faced with the challenge of delivering business value while building capability and simultaneously ramping up the internal team by leveraging new entrants and existing members from other teams. To address this triage of challenges, it is critical to achieve:
- A comprehensive training which focuses on real-world challenges within the business context of Petronas
- Providing the team with custom developed open source tools and utilities which could be directly used for solving key steps in the data science processes
- Guiding data scientists during the execution of a real-life data science project
Actify Data Labs created a custom training program for the Petronas team focusing on the exact needs of the oil and gas industry, using simulated data to mimic actual processes from various processes including sensor data for refinery processes, sales & marketing, distribution and finance. Actify Data Labs used its own proprietary methodology that involved real covariance matrix from the client and copula-based approach for simulating back the actual data. The training also covered the change management, governance and implementation challenges for implementing a data science solution. Methodologies for identifying the right problem and calculating business benefit were the other key aspects of the training.
Actify Data Labs also created custom open source utilities which were integrated into the data science process. These included utilities for data validation, missing imputation, performing cross validation, simulation and making AI solutions explainable. Actify Data Labs also provided a comprehensive feature engineering utility to derive features from streaming sensor data.
Actify Data Scientists helped the Petronas team execute projects and provided guidance as reviewers who could helped them identify the right methodology and perform comprehensive evaluation of alternatives.