Julius Adewopo

Continuous access to up-to-date food price data is crucial for monitoring food security and responding swiftly to emerging risks. However, in many food-insecure countries, price data is often delayed, lacks spatial detail, or is unavailable during crises when markets may become inaccessible, and rising prices can rapidly exacerbate hunger. Recent innovations, such as AI-driven data imputation and crowdsourcing, present new opportunities to generate continuous, localized price data. This paper evaluates the reliability of these approaches by comparing them to traditional enumerator-led data collection in northern Nigeria, a region affected by conflict, food insecurity, and data scarcity. The analysis examines crowdsourced prices for two staple food commodities, maize and rice, submitted daily by volunteers through a smartphone application over 36 months (2019–2021), and compares them with data collected concurrently by trained enumerators during the final eight months of 2021. Additionally, the crowdsourced dataset is compared to AI-imputed prices from the World Bank’s Real-Time Prices (RTP) database. Data from the alternative methods reflected similar price inflation trends during the COVID-19 pandemic. Pearson’s correlation coefficients indicate strong statistical agreement between crowdsourced and enumerator-collected prices (r = 0.94 for yellow and white maize, r = 0.96 for Indian rice, and r = 0.78 for Thailand rice). Furthermore, the crowdsourced data shows a high correlation with the AI-imputed prices (r = 0.99 for maize, and r = 0.94 for rice). The results from additional statistical tests of normality and paired means shows that the discrepancies between price datasets are consistent with measurement error rather than differences in actual price dynamics. Further tests of equivalence confirmed that enumerator and crowdsourced prices represent the same underlying market processes for specific commodity subtypes, and connotes that crowdsourced price data is a credible reference for validating AI-imputed estimates. The results support the use of AI imputation and crowdsourcing methods to improve price data collection and track market dynamics in near real time. These data innovations can be particularly valuable in areas that are underrepresented in national aggregate data due to limited monitoring capacity, and where high-frequency local data can aid targeted interventions.

The fourth agricultural revolution (or Agriculture 4.0) promises to lead the way to an agricultural sector that is smarter, more efficient, and more environmentally and socially responsible. Digital and data generating tools are seen as critical enablers for this transformation and are expected to make farming more planned, predictive, productive, and efficient. To make this vision a reality, agricultural producers will first adopt and use the technologies, but this is easier said than done. Barriers such as limited digital infrastructure, low (digital) literacy, low incomes, and socio-cultural norms are major factors causing sub-optimal access to and use of digital technologies among smallholder farmers. Beyond these use challenges of access and usage, limited evidence exists to support the notion that extant digital technologies add enough value to provide substantial benefits for targeted farmers. In this paper, we unravel insights from a six-year digital agriculture innovation project which was implemented to develop and deploy multi-modal digital tools for the control of a major banana disease. By reaching over 272,200 smallholder farmers in Rwanda through a smartphone app, unstructured supplementary service data, a chatbot, and other ancillary channels, we assessed various assumptions regarding intrinsic motivation, incentives, and skills retention among the target digital tool users. These insights suggest that embedding digital innovation requires intentional user-engagement, proper incentivization of next-users, and targeted communication to foster adoption. We present ten (10) salient, but non-exhaustive, lessons to showcase the realities of developing and delivering digital tools to farmers over an extended period, spanning from ideation, development, and testing to scaling stages. The lessons are relevant for a broad audience, including stakeholders across the digital innovation space who can utilize our experiential notes to guide the development and deployment of similar digital innovations for improved outcomes in smallholder farming systems.

Timely and reliable monitoring of food market prices at high spatial and temporal resolution is essential to understanding market and food security developments and supporting timely policy and decision-making. Mostly, decisions rely on price expectations, which are updated with new information releases. Therefore, increasing the availability and timeliness of price information has become a national and international priority. We present two new datasets in which mobile app-based crowdsourced daily price observations, voluntarily submitted by self-selected participants, are validated in real-time within spatio-temporal markets (pre-processed data). Then, they are reweighted weekly using their geo-location to resemble a formal sample design and allow for more reliable statistical inference (post-sampled data). Using real-time data collected in Nigeria, we assess the accuracy and propose that our reweighted estimates are more accurate with respect to the unweighted version. Results have important implications for governments, food chain actors, researchers and other organisations.

Banana Xanthomonas wilt (BXW) is a major threat to banana production in Rwanda, causing up to 100% yield loss. There are no biological or chemical control measures, and little is known about the potential direction and magnitude of its spread; hence, cultural control efforts are reactive rather than proactive. In this study, we assessed BXW risk under current and projected climates to guide early warning and control by applying the maximum entropy (Maxent) model on 1,022 georeferenced BXW datapoints and 20 environmental variables. We evaluated the significance of variables and mapped potential risk under current and future climates to assess spatial dynamics of the disease distribution. BXW occurrence was reliably predicted (mean validation AUC values ranging from 0.79 to 0.85). Precipitation of the coldest quarter, average maximum monthly temperature, annual precipitation, and elevation were the strongest predictors, which were responsible for 22.1, 13, 12.6, and 9.4% of the observed incidence variability, respectively, while mean temperature of the coldest quarter had the highest gain in isolation. Furthermore, the most susceptible regions (western, northern, and southern Rwanda) were characterized by elevation (1,350 to 2,000 m), annual precipitation (900 to 1,700 mm), and average temperature (14 to 20°C), among other variables, suggesting that a consistent, rainy, and warm climate is more favorable for BXW spread. Under the future climate, the risk was predicted to increase and spread to other regions. We conclude that climate change will likely exacerbate BXW-related losses of banana land area and yield under the influence of temperature and moisture. Our findings support evidence-based targeting of extension service delivery to farmers and national early warning for timely action.

Banana (and plantain, Musa spp.), in sub-Saharan Africa (SSA), is predominantly grown as a mixed crop by smallholder farmers in backyards and small farmlands, typically ranging from 0.2 ha to 3 ha. The crop is affected by several pests and diseases, including the invasive banana bunchy top virus (BBTV, genus Babuvirus), which is emerging as a major threat to banana production in SSA. The BBTV outbreak in West Africa was first recorded in the Benin Republic in 2010 and has spread to the adjoining territories of Nigeria and Togo. Regular surveillance, conducted as part of the containment efforts, requires the identification of banana fields for disease assessment. However, small and fragmented production spread across large areas poses complications for identifying all banana farms using conventional field survey methods, which is also time-consuming and expensive. In this study, we developed a remote sensing approach and machine learning (ML) models that can be used to identify banana fields for targeted BBTV surveillance. We used medium-resolution synthetic aperture radar (SAR), Sentinel 2A satellite imagery, and high-resolution RGB and multispectral aerial imagery from an unmanned aerial vehicle (UAV) to develop an operational banana mapping framework by combining the UAV, SAR, and Sentinel 2A data with the Support Vector Machine (SVM) and Random Forest (RF) machine learning algorithms. The ML algorithms performed comparatively well in classifying the land cover, with a mean overall accuracy (OA) of about 93% and a Kappa coefficient (KC) of 0.89 for the UAV data. The model using fused SAR and Sentinel 2A data gave an OA of 90% and KC of 0.86. The user accuracy (UA) and producer accuracy (PA) for the banana class were 83% and 78%, respectively. The BBTV surveillance teams used the banana mapping framework to identify banana fields in the BBTV-affected southwest Ogun state of Nigeria, which helped in detecting 17 sites with BBTV infection. These findings suggest that the prediction of banana and other crops in the heterogeneous smallholder farming systems is feasible, with the precision necessary to guide BBTV surveillance in large areas in SSA.

Spatially explicit information on soil variability is relevant for agronomic decisions; however, such information is limited in the northern Guinea savanna (NGS) agroecological zone of Nigeria. This study was conducted to delineate soil nutrient management zones (MZs), based on spatial variability of soils in the smallholder maize-based farming system within the NGS. Two hundred and eighty-nine soil samples were analyzed for some physical and chemical properties. Principal component analysis (PCA) was used to aggregate the soil properties into four principal components, which accounted for about 60 of the variation in the data, and spatial variability was assessed with a semivariogram. The ordinary kriging technique was used to predict soil properties at unsampled locations, while weighted overlay analysis was conducted to delineate nutrient management zones. Results showed that total nitrogen (0.06), available phosphorus (5.6mgkg⁻¹), organic carbon (0.66), and effective cation exchange capacity (5.6cmol₍₊₎ kg⁻¹) are below optimal requirement for maize production. Four MZs were identifiable in the region with the highest fertility (MZ3 and MZ4) associated with the northern area but covering a relatively small part (9.1). The differences observed in soil properties among the MZs suggest that each zone requires different agronomic management, especially in relation to fertilizer application.

This chapter provides the first comprehensive review of digital tools and technologies available for the identification, monitoring, and control of pests and diseases, with an emphasis on root, tuber, and banana (RT&B) crops. These tools include systems based on identification keys, human and artificial intelligence-based identification based on smart applications, web interfaces, short messages services (SMS), or combinations thereof. We also present ideas on the use of image recognition from smartphones or unmanned aerial vehicles (UAVs) for pest and disease monitoring and data processing for modeling, predictions, and forecasting regarding climate change. These topics will be presented in the context of their current development and future potential but also the challenges, limitations, and innovative approaches taken to reach end users, particularly smallholder farmers, and achieve impacts at scale. Finally, the scope and limitation of private sector involvement demonstrates the need of publicly funded initiatives to maximize sharing of data and resources to ensure sustainability of unbiased advice to farmers through information and communication technology (ICT) systems.

Crop monitoring is crucial to understand crop production changes, agronomic practice decision-support, pests/diseases mitigation, and developing climate change adaptation strategies. Banana, an important staple food and cash crop in East Africa, is threatened by Banana Xanthomonas Wilt (BXW) disease. Yet, there is no up-to-date information about the spatial distribution and extent of banana lands, especially in Rwanda, where banana plays a key role in food security and livelihood. Therefore, delineation of banana-cultivated lands is important to prioritize resource allocation for optimal productivity. We mapped the spatial extent of smallholder banana farmlands by acquiring and processing high-resolution (25 cm/px) multispectral unmanned aerial vehicles (UAV) imageries, across four villages in Rwanda. Georeferenced ground-truth data on different land cover classes were combined with reflectance data and vegetation indices (NDVI, GNDVI, and EVI2) and compared using pixel-based supervised multi-classifiers (support vector models-SVM, classification and regression trees-CART, and random forest–RF), based on varying ground-truth data richness. Results show that RF consistently outperformed other classifiers regardless of data richness, with overall accuracy above 95%, producer’s/user’s accuracies above 92%, and kappa coefficient above 0.94. Estimated banana farmland areal coverage provides concrete baseline for extension-delivery efforts in terms of targeting banana farmers relative to their scale of production, and highlights opportunity to combine UAV-derived data with machine-learning methods for rapid landcover classification.

The rapid assessment of maize yields in a smallholder farming system is important for understanding its spatial and temporal variability and for timely agronomic decision-support. We assessed the predictability of maize grain yield using unmanned aerial/air vehicle (UAV)-derived vegetation indices (VI) with (out) biophysical variables on smallholder farms. High-resolution imageries were acquired with UAV-borne multispectral sensor at four and eight weeks after sowing (WAS) on 31 farmer managed fields (FMFs) and 12 nearby nutrient omission trials (NOTs) sown with two genotypes (hybrid and open-pollinated maize) across five locations within the core maize region of Nigeria. Acquired multispectral imageries were post-processed into three VIs, normalized difference VI (NDVI), normalized difference red-edge (NDRE), and green-normalized difference VI (GNDVI) while plant height (Ht) and percent canopy cover (CC) were measured within georeferenced plot locations. Result shows that the nutrient status had a significant effect on the grain yield (and variability) in NOTs, with a maximum grain yield of 9.3 t/ha, compared to 5.4 t/ha in FMFs. Generally, there was no relationship between UAV-derived VIs and grain yield at 4WAS (r < 0.02, p > 0.1), but significant correlations were observed at 8WAS (r ≤ 0.3; p < 0.001). Ht was positively correlated with grain yield at 4WAS (r = 0.5, R² = 0.25, p < 0.001) and more strongly at 8WAS (r = 0.7, R² = 0.55, p < 0.001), while the relationship between CC and yield was only significant at 8WAS. By accounting for within- and between-field variations in NOTs and FMFs (separately), predictability of grain yield from UAV-derived VIs was generally low (R² ≤ 0.24); however, the inclusion of ground-measured biophysical variable (mainly Ht) improved the explained yield variability (R² ≥ 0.62, Root Mean Square Error of Prediction, RMSEP ≤ 0.35) in NOTs but not in FMFs. We conclude that yield prediction with UAV-acquired imageries (before harvest) is more reliable under controlled experimental conditions (NOTs), compared to actual farmer managed fields where various confounding agronomic factors can amplify noise-signal ratio.