Spatiotemporal evolution of Nigeria's armed conflicts and terrorism and the associated shift in social perceptions

Nigeria, with a population exceeding 200 million and diverse ethnic composition, remains among the most severely affected African countries by armed conflict and terrorism. Despite government claims that Boko Haram attacks in the Northeast have declined recently, overall yearly armed conflict events (battles, explosions/remote violence, riots, and violence against civilians) have remained high. ACLED recorded over 4500 conflicts in both 2021 and 2022, each with more than 10,000 fatalities, yet international attention has been limited relative to other conflicts such as the Russia–Ukraine war. Existing research spans tourism effects, regime variation, trust, modeling, and prediction of terrorism, and identifies potential causes including poverty, education, failed states, minority discrimination, and religion. Assessments by policy institutions underscore corruption, injustice, and marginalization as drivers in the Sahel. However, these insights have not sufficed to guide concrete measures in Nigeria. This study aims to systematically characterize Nigeria’s spatiotemporal conflict evolution via ACLED and to construct media-based indices from GDELT to assess changes in social perception regarding conflict and terrorism.

Prior work has examined diverse facets of terrorism: impacts on tourism (Drakos and Kutan, 2003), regime-related variations (Findley and Young, 2011), terrorism and trust (Blomberg et al., 2011), modeling approaches (Brandt and Sandler, 2010; Clauset and Wiegel, 2010; D'Orsogna and Perc, 2015; Gao et al., 2017; Helbing et al., 2015; Python et al., 2019), and forecasting (Blair et al., 2017; Cederman and Weidmann, 2017; Hegre et al., 2013, 2019, 2021a,b; Python et al., 2021; Weidmann and Ward, 2010; Witmer et al., 2017). For Africa and Nigeria, studies link economics, governance, military expenditure, and terrorism (Abadie, 2006; Abid and Sekrafi, 2020; Blomberg et al., 2004, 2007; Emeka et al., 2024; Keefer and Norman, 2008; Ogbuabor et al., 2023). Identified causes include poverty and education (Krueger and Malečková, 2002), failed states (Piazza, 2008), minority discrimination (Piazza, 2012), and religion (Jones, 2006). Dim (2017) proposed integrating poverty, relative deprivation, and social identity theories for Boko Haram’s persistence. Policy assessments (Chatham House; UN Security Council briefings) emphasize corruption, unpaid troops, militaries perceived as oppressors, and broader grievances (despair, poverty, hunger, lack of services, unemployment, unconstitutional regime changes) enabling jihadist insurgencies and dissemination of violent ideologies online. The paper positions its contribution as providing quantitative, readily computable indices to better guide practical interventions in Nigeria.

Data: ACLED conflict events and fatalities for Nigeria from January 2013 to December 2023 were used, focusing on four event types—battles, explosions/remote violence, riots, and violence against civilians (strategic development and protests excluded due to sparse data). Each ACLED record includes event type, date, location, actors, notes, and fatalities. GDELT events (1979–present) with Goldstein Scale values (−10 to 10) were used to derive indices of cooperative and conflicting national activity. Democracy Index (Economist Intelligence Unit) and Corruption Perceptions Index (Transparency International) were added as potential determinants; data from 2016–2022 were used. Analytical framework: (1) Distributional analysis: Examine whether daily fatalities follow lognormal distributions by testing normality of log-transformed nonzero fatalities within each year. (2) Shannon entropy: Spatial entropy of conflicts/fatalities computed by partitioning Nigeria into 0.5°×0.5° grid cells, calculating pk as the share of events (or fatalities) per cell, and computing ENT = −Σ pk log2 pk. Temporal entropy of fatalities is computed monthly by binning log-transformed nonzero fatalities and calculating entropy of the monthly distribution to assess uniformity over time. (3) Revealed Comparative Wealth (RCW): RCW(k,t) = [W(k,t)/popul(k,t)] / [ΣW(k,t)/Σpopul(k,t)], with W approximated by nominal GDP; captures a country’s GDP per capita relative to the world average, removing nonstationarity and enabling comparisons across time and countries. (4) Statistical analysis: Pearson correlation and R² to evaluate associations between annual totals of events/fatalities and RCW, democracy, and CPI (primarily 2016 onward); note the short series precludes formal regression. (5) Media-derived indices: Define Bilateral Relationship (BLR) indices using counts and average Goldstein scales for cooperative (+) and conflict (−) events, with weighting of partner importance via Goldstein-weighted shares. Aggregate to National Activity Index (NAI±): the ratio of total positive (or negative) Goldstein-weighted interactions to the sum of positive and negative totals for a given country. Trends for NAI± and fatality series were extracted using an adaptive filtering algorithm demonstrated to be near-optimal (Gao et al., 2009, 2011). Comparative spatial mapping was conducted for 2014, 2016, and 2021 to represent distinct conflict phases.

- 2014–2015 exhibited extreme fatalities, particularly in battles, explosions/remote violence, and violence against civilians; e.g., on 1 Feb 2015 battle-related deaths exceeded 500, and the Baga massacre week included ~800 deaths on 7 Jan 2015. Riots caused far fewer deaths. Since 2016, deaths from violence against civilians decreased relative to earlier peaks. - Monthly mean fatalities were high in 2014–2015 and have slowly decreased since; monthly standard deviation showed a similar pattern until 2018, then fluctuated modestly, increasing variability after 2022. - Monthly total conflict events increased steadily from 2013, peaking around end-2020 and remaining high thereafter. Monthly fatalities showed a similar rising pattern since 2016; earlier, totals diverged due to few high-fatality events in 2014–2015. - Annual fatalities within each year are approximately lognormal: PDFs of log10 fatalities are bell-shaped across years; since 2016 these PDFs shift toward lower values, consistent with decreasing mean fatalities. - Temporal entropy (monthly fatalities entropy) has been steadily decreasing since 2016, especially after COVID-19, indicating increasing temporal uniformity of fatal outcomes even as event counts remained high. - Spatial spread increased markedly post-COVID-19: maps show expansion from concentrated Northeast hotspots (2014, 2016) to nationwide distribution by 2021. Quantitatively, the number of 0.5° cells with conflicts/fatalities rose from a few hundred (2016) to nearly 1000 (2021). Shannon entropy of spatial distributions (events and fatalities) increased notably after the pandemic’s onset, especially for battles and violence against civilians. - Concrete totals: 2016 recorded 1408 events and 4896 fatalities; 2021 recorded 4546 events and 10,880 fatalities (events >3× 2016; fatalities >2× 2016). ACLED totals exceeded 4500 conflicts and 10,000 fatalities in both 2021 and 2022. - Media-based perceptions: NAI− (conflict) and NAI+ (cooperation) show overall upward trends from 2012, reflecting increased conflict salience in media. Short-term surges in NAI− align with major events (e.g., 2014–2015 Boko Haram violence; multiple kidnappings and insurgent actions in 2021–2022). From the second half of 2021, NAI− declines notably despite sustained conflict levels, suggesting normalization (reduced media/public sensitivity). - Correlations with conflict severity: NAI− trend correlates strongly with the battles fatality trend (r ≈ 0.667; R² ≈ 0.45) and overall fatalities (r ≈ 0.634). Other event-type fatality trends show weak or no correlation with NAI−, implying battles dominate media attention. - Socioeconomic and governance linkage: Post-2016—particularly post-COVID-19—declines in RCW (GDP per capita relative to world average), democracy index, and CPI coincide with increased conflict fatalities/events. RCW fell from ~0.3 (2014) to ~0.17 (2022). Scatter plots (2016–2022) show strong associations: R² values reported up to ~0.545 (fatalities vs RCW), ~0.575 (events vs democracy), and ~0.912–0.961 (fatalities/events vs CPI), indicating tight relationships between worsening economic/governance indicators and higher conflict intensity.

The study addresses the central question of how Nigeria’s armed conflicts and terrorism have evolved spatiotemporally and how societal perception has shifted. By combining ACLED conflict data with information-theoretic measures, the analysis shows a dual dynamic: spatial diffusion (higher spatial entropy) and temporal regularization of fatal outcomes (lower temporal entropy), especially after COVID-19. The nationwide expansion of conflict activity by 2021 underscores a broadening security challenge beyond traditional hotspots. Media-derived indices from GDELT reveal that public/media attention (proxied by NAI−) tracks intense conflict periods and battle-driven fatalities but, from late 2021 onward, declines despite persistently high conflict levels, indicating societal adaptation to a “new normal.” This perceptual shift has implications for domestic and international responsiveness and resource mobilization. The strong associations between increased conflict severity and declines in RCW, democracy, and CPI support the interpretation that economic decay, democratic erosion, and corruption are closely linked to conflict intensification and spread. Policy relevance follows: conflict mitigation requires not only security responses and regional/international cooperation but also sustained economic revitalization, improved governance, and anti-corruption measures. The methodology—entropy metrics, RCW, and NAI—offers a practical toolkit to monitor evolving risk and societal perception in near real time.

This work contributes a quantitative, index-based framework to characterize Nigeria’s conflict dynamics and societal perception. Key findings are that conflicts have become more spatially widespread and temporally uniform; that post-COVID-19 patterns reflect accelerated diffusion; and that economic decay (declining RCW), erosion of democracy, and rising corruption are strongly associated with increased conflict intensity. The newly constructed National Activity Index (NAI±) from GDELT serves as a proxy for societal attention: NAI− rose markedly around major conflict surges but declined from late 2021 despite continued high conflict levels, indicating normalization of severe violence in public/media perception. Policy implications include the need for comprehensive strategies coupling security and governance reforms with economic support—potentially via upgraded industrial chains and enhanced international assistance—to address structural constraints highlighted by global economic hierarchy. Future research should extend time horizons to enable formal causal inference, integrate additional socioeconomic and institutional variables, and investigate the mental health burden in highly affected African contexts, where impacts may exceed those observed in prior Western-focused studies.

- Short time series for RCW, democracy index, and CPI (primarily 2016–2022) preclude robust formal regression and causal inference; analyses rely on correlations and scatter plots with trend extraction. - Two ACLED categories (strategic development and protests) were excluded due to limited data, potentially omitting relevant dynamics. - Media-derived indices (GDELT) reflect coverage and salience, which can be influenced by reporting biases and evolving media ecosystems; NAI captures perception proxies rather than ground truth. - Entropy-based measures summarize dispersion/uniformity but do not specify underlying causal mechanisms. - Spatial resolution (0.5° grid) may blur finer-grained local heterogeneities.