About Me
Data scientist with an MPhil in Computer Science, specialising in machine learning and graph-based anomaly detection in large billing datasets. Designed and developed the Billing Anomaly Detection System (BDAS) in Python…
Data scientist with an MPhil in Computer Science, specialising in machine learning and graph-based anomaly detection in large billing datasets. Designed and developed the Billing Anomaly Detection System (BDAS) in Python and SQL, which flags suspected meter tampering, abnormal consumption, prolonged estimated billing and under-billing: a revenue-assurance problem with clear parallels to anomaly and fraud-risk detection on payment transactions. Works with billing, collection and customer account data at Ghana Water Limited (GWL), a national utility with a metered, subscription-based customer base, and has turned analysis into measurable results, including reducing Non-Revenue Water from above 50% to approximately 21% across approximately 4,989 accounts. Combines hands-on Python, SQL, Jupyter and Azure skills with experience communicating analytical findings to management and operational teams.
Experience
Officer, ICT
<p>Apply AI/ML techniques to revenue intelligence, including the Billing Anomaly Detection System (BDAS), to detect revenue leakage and support data-driven decision-making.</p><p> Build data workflows and digital applications using Python, SQL and Microsoft Azure Functions, and prototype tools with modern low-code platforms (Lovable).</p><p> Develop Power BI dashboards and reports that communicate revenue, customer and operational performance to management.</p><p> Support the enterprise reporting and eBilling systems that generate the billing data used for analysis.</p>
District Manager
<p>Used billing, collection and operational KPIs to direct field and commercial activity, reducing Non-Revenue Water from above 50% to approximately 21%.</p><p> Achieved billing and collection performance above 105% of target.</p><p> Managed approximately 4,989 customer accounts and a team of 16 staff, building direct operational knowledge of customer billing and payment behaviour.</p>
Regional ICT Officer
<p>Supported the systems and users underpinning billing, reporting and day-to-day operations at regional and district levels.</p><p> Served in a regional ICT leadership capacity, supporting ICT operations and technology management.</p>
PROJECTS
Billing Anomaly Detection System
<p>Designer and developer of BDAS,</p><ul><li>defined the business problem: revenue leakage from meter tampering, abnormal consumption, prolonged estimated billing and under-billing</li><li>researched graph-based machine-learning approaches and related techniques for billing anomaly detection</li><li>built the system in Python and SQL so it works on large billing datasets</li><li>applied it at Ghana Water Limited as part of your AI/ML revenue-intelligence work, to support data-driven decisions</li></ul><p>Tools and techniques</p><ul><li>Python: data analysis and machine-learning development</li><li>SQL: extracting and manipulating billing data</li><li>Jupyter Notebook and Google Colab: development and experimentation</li><li>Graph Neural Networks: </li><li>Anomaly detection and related machine-learning methods</li></ul><p></p><p>Achievements</p><ul><li>Turned a manual, hard-to-spot problem into an automated one: BDAS flags suspected anomalous accounts for investigation.</li><li>Gave revenue teams a prioritised, data-driven view of potential revenue leakage.</li><li>Covers four types of anomaly in one system: meter tampering, abnormal consumption, prolonged estimated billing and under-billing.</li><li>Grounded in postgraduate research, applying recent graph-based models to a real utility billing problem.</li></ul><p></p>