As part of my ongoing work in data science, I performed an exploratory analysis of network traffic data, focusing on feature distribution, normalization, and correlation analysis.
I examined the distributions of dst_host_srv_count and srv_serror_rate to identify skewness, spread, and variations in network traffic characteristics.
I applied Min-Max scaling to transform numerical features into a standardized [0, 1] range for consistent preprocessing.
I investigated the distribution of num_root to understand the frequency and behaviour of discrete observations.
I developed a feature correlation heatmap to examine linear relationships and assess potential feature dependencies.
- My analysis identified differences in feature distributions and scales, observed highly skewed distributions in certain network traffic variables, and examined the correlation structure to better understand feature relationships.
I highlighted the importance of preprocessing and statistical analysis before model development. Data preprocessing is not simply a transformation step. It is a critical part of building reliable machine learning pipelines and making informed modelling decisions.
I am continuing to strengthen my understanding of statistical analysis, feature engineering, and practical machine learning workflows.