2024 Project Review and documentation of a project carried out during my internship at HNG TECH focused on data cleaning and anomaly detection in polling-unit-level electoral datasets. I returned to the work to organize the methodology and findings to showcase real analytical impact in my professional portfolio.
Why it matters Electoral credibility is key to democracy. Using Python, data per polling unit in Zamfara State was analyzed to identify statistical anomalies that may indicate irregularities, transcription errors, or procedural failures. It is a lightweight and reproducible approach to electoral auditing without the need for GIS tools like ArcGIS or QGIS, only Python and statistical thinking.
Data source and preparation We worked with a CSV file that includes geospatial coordinates, registered voters, accredited voters, votes per party, and polling unit metadata. The coordinates were cleaned and organized manually in Excel before the analysis. The workflow was reproducible using Pandas, NumPy, and Matplotlib and executable in Google Colab for cloud reproducibility.
Anomaly detection Outlier scores were calculated using z-scores and domain-based heuristic rules. Polling units were ranked by anomaly severity per party, allowing audits to be prioritized. The method is scalable and can be applied to other states or electoral processes without GIS dependency.
Key findings Units PU 108 and PU 104 in Birnin Magaji showed high atypical scores for APC and PDP. PU 321 and PU 124 in Tsafe presented negative counts in transcriptions indicating transcription errors. Geospatial clustering of anomalies was observed in certain local districts and cross-party deviations, suggesting systematic issues rather than isolated cases.
Technical aspects Implementation with pure Python, without ArcGIS or QGIS, using Pandas, NumPy, and Matplotlib. Execution in Google Colab for collaboration and reproducibility. The approach allows integrating anomaly detection into electoral result verification processes and scaling to other data analysis contexts.
Recommendations Audit the flagged polling units, correct metadata gaps and transcription errors, and integrate automatic anomaly detection into official result verification. Establish data pipelines with quality controls and alerts to prevent manual errors from reaching final reports.
Professional impact The project demonstrates how analytics and applied artificial intelligence with good practices can bring transparency to politically sensitive processes. It is also an example of how to build reproducible and scalable solutions that add value in audits and data quality controls.
About Q2BSTUDIO Q2BSTUDIO is a software development company specialized in custom applications and bespoke software. We offer artificial intelligence and AI solutions for businesses, AI agents, cybersecurity services, AWS and Azure cloud services, business intelligence and visualization services with Power BI. We develop custom applications that integrate artificial intelligence models, security by design, and cloud deployments to ensure performance and compliance.
Featured services At Q2BSTUDIO we design custom software, implement artificial intelligence solutions for businesses, build AI agents, and offer consulting in cybersecurity and AWS and Azure cloud services. We also deliver business intelligence solutions with Power BI to facilitate data-driven decision-making.
Access the full analysis To view the complete notebook with code, tables, and transcription breakdown, visit https://www.linkedin.com/posts/rahmah-abubakar-243058288_analysis-process-activity-7253521166152163328-zh2n?utm_source=share&utm_medium=member_android&utm_rcm=ACoAAEXGTFMBU8WaWDL8Z4Wl7HzBBVCiQx_eYO4
Conclusion It is not necessary to use ArcGIS or QGIS to solve relevant problems. With a data-centric approach, Python, and a reproducible architecture, anomalies can be detected and prioritized with real impact. If you are looking to develop custom solutions that combine artificial intelligence, cybersecurity, AWS and Azure cloud services, or dashboards with Power BI, Q2BSTUDIO offers comprehensive expertise to turn data into decisions.
