Browsing by Author "Muriithi, Betsy"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Publication
    Enhancing Agricultural Support for Small Scale Farmers in Kenya: An IoT-based Mini Weather Station as a Machine Learning Data Collector
    (IEEE Explore, 2023-12) Itotia, Solomon; Muriithi, Betsy; Gitahi, Stephen; Korir, Phylis; Murigi, Morris; Kimutai, Ronny; Olukuru, John; Sevilla, Joseph
    This paper presents a mini weather station device as a data collector for a machine learning model, aiming to support the Kenyan agricultural sector through small-scale farmers. Most farmers in Kenya practice small-scale farming and often face challenges in accessing timely and accurate weather information, which deters them from making informed decisions on what kind of crops to grow and the type of resources to allocate to their farms. We propose designing and deploying an affordable IoT mini weather station that collects real-time weather data to address this issue. The device has sensors that collect meteorological parameters such as humidity, temperature, light intensity, and atmospheric pressure. The collected data is transmitted to a cloud server and can be used as input for AI-powered machine learning models for forecasting and advisory systems to personalize recommendations to farmers, such as optimal planting time, irrigation schedules, and pest management strategies based on the prevailing weather.
  • Thumbnail Image
    Publication
    Enhancing Food Security in Africa with a Predictive Early Warning System on Extreme Weather Phenomena
    (Research Square, 2022-03-02) Igobwa, Alvin M.; Gachanja, Jeremy; Muriithi, Betsy; Olukuru, John; Wairegi, Angeline Rehema; Rutenberg, Isaac
    Climate change is predicted to exacerbate Africa’s, already, precarious food security. Climate models, by accurately forecasting future weather events, can be a critical tool in developing countermeasures to reduce crop loss, decrease adverse effects on animal husbandry and fishing, and even help insurance companies determine risk for agricultural insurance policies – a measure of risk reduction in the agricultural sector that is gaining prominence. In this paper, we investigate the efficacy of various open-source climate change models and weather datasets in predicting drought and flood weather patterns in northern and western Kenya and discuss practical applications of these tools in the country’s agricultural insurance sector. We identified two models that may be used to predict flood and drought events in these regions. The combination of Artificial Neural Networks (ANNs) and weather station data was the most effective in predicting future drought occurrences in Turkana and Wajir with accuracies ranging from 78% to 90%. In the case of flood forecasting, Isolation Forests models using weather station data had the best overall performance. The above models and datasets may form the basis of a more objective and accurate underwriting process for agricultural index-based insurance, as we expound in the paper.