Investigating Nigerian Election Integrity

Introduction

In the aftermath of Nigeria's recent election, widespread irregularities have sparked legal challenges against the Independent National Electoral Commission (INEC). This project focuses on scrutinizing the results from Delta State, specifically targeting outlier detection to enhance transparency and address concerns over electoral integrity.

Objective

The primary objective of this analysis is to uncover potential voting irregularities and ensure the transparency of the election results by identifying outlier polling units in Delta State, Nigeria. These outliers are characterized by voting results that deviate significantly from neighbouring units, which might indicate undue influences or rigging.

Methodology:

Data Cleaning: The dataset was first cleaned to ensure consistency.

Geospatial Analysis: Longitude and latitude values were integrated into each polling unit or ward using the OpenCage Geocoding API. Additionally, Google Sheets add-ons were employed for data manipulation and to facilitate the inclusion of geographical coordinates.

Outlier Detection: The Isolation Forest algorithm was used to calculate outlier scores for each party in each polling unit. This algorithm helps identify data points that are significantly different from the rest.

Outliers Per Party

APC

High Outlier Scores: APC has notably high outlier scores in several polling units , such as "AZONOBA P/S - UBULUBU I, UBULUBU" with a score of 341.14. This suggests that APC may have received significantly more votes than expected in these areas, potentially indicating strong local support or irregularities favoring APC.

PDP :Moderate Outlier Scores: PDP also shows outlier scores that indicate higher-than-expected vote counts in some PU-Names, like "KANIDILIMA P/S - ISHIEKPE QUARTERS, ONICHA-UKU" with a score of 18.50. This could also signify localized support for PDP or anomalies that benefited the party.LP (Labor Party):High Outlier Scores: LP had extremely high outlier scores, reaching up to 99.39 in "KANIDILIMA P/S - ISHIEKPE QUARTERS, ONICHA-UKU" and "KANIDILIMA P/S - OGBE AKWU, ONICHA-UKU". Such scores suggest LP received disproportionately high votes compared to the norm, indicating potential anomalies or strong local support.NNPP (New Nigeria People's Party):Varied Outlier Scores: NNPP generally shows lower outlier scores compared to APC and LP, with scores like 3.25 in "AZONOBA P/S - UBULUBU III, UBULUBU". These scores suggest moderate deviations from expected vote counts, possibly reflecting varying levels of support or minor anomalies.

Detailed Examples of the Top 3 Outliers

ZAPPA MIXED SEC. SCH. I, UMUAGULA/AKWUOFOR PHASE II (LP)

Outlier Score: 574.0

This polling unit shows an extremely high outlier score for Labour Party (LP). The score suggests that the voting pattern here significantly deviates from the average, which indicates potential anomalies or this could also be a unique voter behavior.

GOVERNMENT DAY CARE NURSERY AND PRIMARY SCHOOL (LP)

Outlier Score: 574.0

Similar to the first pooling unit, this polling unit has an identical high outlier score for the Labour Party. This indicates either a data anomaly or a significant deviation in voter preferences.

AWELLE MARKET SPACE (LP)

Outlier Score: 409.0

This unit also shows a very high outlier score for the Labour Party. The score is lower than the top two but still significantly higher than average, suggesting potential irregularities or unique local dynamics.

Summary of Findings:

The Labour Party (LP) shows the highest outlier scores in several polling units, suggesting either voter behavior anomalies or data irregularities.

The APC and PDP parties have fewer and lower outlier scores, indicating more consistent voting patterns across the polling units.

Detailed analysis of the top 3 outliers reveals specific polling units where the Labour Party’s outlier scores are significantly higher than average, warranting further investigation.

Conclusion

Key Insights

1. High Outlier Scores for LP:
   • The Labour Party (LP) has the highest outlier scores in several polling units, indicating potential irregularities or unique voter behaviors in those areas.
   • The top three outliers (ZAPPA MIXED SEC. SCH. I, GOVERNMENT DAY CARE NURSERY AND PRIMARY SCHOOL, and AWELLE MARKET SPACE) all belong to the Labour Party, suggesting a pattern that merits further investigation.
2. Consistent Patterns for Other Parties:
   • The APC and PDP parties had fewer and lower outlier scores, which suggests more consistent voting patterns across different polling units.
   • This consistency could indicate either a stable voter base or voters largely prefer Labour party