I have a Google Colab notebook (link above) where I am combining numerical and textual features in a single exploratory study. The core objective is to drill into the numeric columns with quantile-based techniques—building custom percentiles, detecting outliers, and comparing distributions—while also keeping the surrounding text fields available for cross-referencing and basic summaries. The notebook already loads the data and does some preliminary cleaning, but the workflow needs to be tightened, documented, and extended. Specifically, I want: • Robust quantile calculations and visualisations (pandas, NumPy, seaborn / matplotlib) that highlight how each percentile band behaves. • Clean, reusable functions so I can re-run the analysis on future datasets without rewriting code. • A concise text report generated inside the notebook that pairs the numeric findings with quick text-based context (for example highlighting which product names or comment snippets sit in the extreme quantiles). Please keep everything inside the existing Colab, add clear markdown explanations, and leave cells lightweight so they run within the free tier. A short “next steps” cell at the end outlining any additional analyses you recommend will round things off.