This article reviews the application of machine learning (ML) methods in Internet of Things (IoT) data analysis, focusing on smart cities as a primary use case. The authors aim to assess different ML techniques that address the challenges posed by IoT data, which is characterized by high volume, velocity, variety, and varying quality. The key contributions of the study include: 1. **Taxonomy of ML Algorithms**: The article presents a taxonomy of ML algorithms, explaining how different techniques are applied to extract higher-level information from IoT data. 2. **IoT Data Characteristics**: It discusses the characteristics of IoT data, such as continuous generation, dynamic nature, and varying quality, and how these characteristics impact data processing. 3. **Smart City Use Case**: Smart cities are selected as a primary use case due to their extensive application in IoT, the availability of open datasets, and the potential for improving urban services and quality of life. 4. **Support Vector Machine (SVM) Application**: A detailed exploration of applying SVM to Aarhus Smart City traffic data is provided to predict traffic hours during the day. The article also reviews related literature, discusses IoT applications, communication protocols, computing frameworks, and the challenges of data processing and preparation. It highlights the importance of data quality, semantic data annotation, and the integration of heterogeneous data sources. The taxonomy of ML algorithms is organized into supervised, unsupervised, and reinforcement learning categories, with specific algorithms like K-Nearest Neighbors, Naive Bayes, SVM, linear regression, and clustering methods discussed in detail. The article concludes with a discussion on future research trends and open issues in IoT data analytics.This article reviews the application of machine learning (ML) methods in Internet of Things (IoT) data analysis, focusing on smart cities as a primary use case. The authors aim to assess different ML techniques that address the challenges posed by IoT data, which is characterized by high volume, velocity, variety, and varying quality. The key contributions of the study include: 1. **Taxonomy of ML Algorithms**: The article presents a taxonomy of ML algorithms, explaining how different techniques are applied to extract higher-level information from IoT data. 2. **IoT Data Characteristics**: It discusses the characteristics of IoT data, such as continuous generation, dynamic nature, and varying quality, and how these characteristics impact data processing. 3. **Smart City Use Case**: Smart cities are selected as a primary use case due to their extensive application in IoT, the availability of open datasets, and the potential for improving urban services and quality of life. 4. **Support Vector Machine (SVM) Application**: A detailed exploration of applying SVM to Aarhus Smart City traffic data is provided to predict traffic hours during the day. The article also reviews related literature, discusses IoT applications, communication protocols, computing frameworks, and the challenges of data processing and preparation. It highlights the importance of data quality, semantic data annotation, and the integration of heterogeneous data sources. The taxonomy of ML algorithms is organized into supervised, unsupervised, and reinforcement learning categories, with specific algorithms like K-Nearest Neighbors, Naive Bayes, SVM, linear regression, and clustering methods discussed in detail. The article concludes with a discussion on future research trends and open issues in IoT data analytics.