This research presents an empirical evaluation of CPython and Mojo with respect to two key dimensions: execution performance and qualitative software quality attributes, most notably code readability. While CPython is renowned for its expressive syntax and Mojo for its explicit memory control, the impact of code readability on performance and overall code quality remains unclear. To investigate this, we conduct a systematic empirical comparison using a dual benchmarking methodology that integrates macro-benchmarking (wall time, CPU time) with micro-benchmarking at the function level. Six representative algorithmic workloads are evaluated: random integer generation, Fibonacci computation, matrix multiplication, quick sort, A* pathfinding, and Conway's Game of Life. The study evaluates performance through a combination of quantitative complexity metrics and a two-stage code review process. In terms of performance, Mojo significantly outperforms CPython in numerically intensive and compute-bound tasks, achieving speed-ups of up to 77× in matrix multiplication and notable performance improvements in sorting benchmarks. In contrast, CPython consistently ranks higher in readability and developer experience, offering clearer error messages and more concise implementations. The result confirms Mojo as a promising option for performance-critical components, particularly in scientific and data-intensive computing such as machine learning. Furthermore , a hybrid approach integrating Mojo for computational hotspots within Python-based systems can emerge as the most practical development strategy. The research contributes to a reproducible benchmarking and evaluation framework to support future research as Mojo evolves.

Advancements in fitness tracking systems have revolutionised cycling by making health fitness related data more accessible. The cycling power is an essential indicator when it comes to actual measurement of absolute efforts and performance of athletes and coaches. Unfortunately, cycling power requires a specialist power meter sensor to measure, which adds significant cost to an already expensive sport. This study investigates the potential of machine learning techniques to predict average cycling power without the need for a power meter. Unlike traditional physics-based models, which rely on fixed equations, regression models can learn complex, data-driven patterns to improve accuracy and generalisability. Five regression models — multiple linear regression, random forest regression, XGBoost regression, support vector regression, and bayesian ridge regression were trained on two distinct yet comparable datasets. One dataset was from an online source and the other sourced from cycling clubs local to city of Edinburgh, Scotland. Feature selection and hyper-parameter tuning were performed to optimise each model. Support vector regression model, which has previously missing from literature in similar applications, emerged as the best performing model — achieving a mean R2score of 0.91. These results advance the field of cycling power prediction by identifying a more accurate regression model while also providing a comparative analysis with other regression techniques.

Google Lighthouse is a tool made by Google for auditing web pages performance, accessibility, SEO, and best practices with the intention of improving the quality of the websites. This allows software developers to understand areas of improvement within a website. However, accessibility becomes an issue when it comes to rapidly handling multiple input for generating performance report consisting different formats with better accuracy. In this work, a test on set of improvements to enhance the reference Google Lighthouse architecture was conducted to enable it to process multiple file input types with the focus of producing different output format including better accuracy. Multiple experimental runs involving different size of file input were conducted by comparing the results from using the modified Lighthouse with the original Lighthouse. The results proves the feasibility of the enhancement and can further improve the accuracy of performance audit report with known Lighthouse performance metrics.

Semantic technology can provide a bridge between smart applications and Internet of Things (IoT) to enable possible integration and interoperability of data produced by heterogeneous devices. In IoT, data quality plays an important role when it comes to interfacing sensor readings with real-time applications at the basic atomic level. Popular techniques of machine learning and point-based calibrations are inadequate due to inability to perform semantic reasoning and interoperability on sensor streams even in real time. In this paper, a layered software framework based on semantic technologies is developed to maintain the consistency of data streams produced by physical sensors that interprets measurements as numeric values. The framework shows how semantic modelling and reasoning can be applied to validate the consistency of data streams while placing emphasis on the temporal characteristics of the stream. The evaluation of the approach involves analysing the effects of different Resource Description Format(RDF) data serializations on the response times of the reasoning engine and throughput of continuous semantic stream query execution. The outcome of experiments indicates the semantic framework as a promising approach for stream validation in Smart Spaces and other related IoT domains.

Smart Spaces currently benefits from Internet of Things (IoT) infrastructures in order to realise its objective. In many cases, it demonstrates this through certain automated applications that relies on sensor streams that comes with some uncertainties in measurements. However, these sensor data tend to be uncertain or fault-prone due to the faults of the sensor either themselves or the wireless sensor networks. Sometimes, the extreme operating condition of the sensor can be a contributing factor to the uncertainty. The proposed approach provides a software framework that aims at homogenising, annotating and reasoning over these data. The framework consists of four layers that utilizes the semantic process involving a domain ontology and reasoning process to deliver improved quality data streams to applications. This will allow for early detection of missing data points and enhancing the accuracy of decisions and actions in such spaces.