Data visualizations summarize vast amounts of complex data into clear, actionable formats. Choosing the right visualization can be challenging—especially as automated insights, dynamic embedded reporting, and Generative AI shift how we interact with dashboards. This guide outlines how to select the right chart based on your analytical goal, updated for modern business intelligence and the AI-driven data stack. The Modern Visualization Engine: How AI Has Changed Charting AI tools (such as ThoughtSpot, Tableau Einstein, Power BI Copilot, and LLM-driven Python interpreters) now automate chart creation via natural language prompts like "Show me ARR growth by region over time." However, AI models often fall back on default options (like basic pie charts or crowded line plots). Knowing why a visualization works remains essential for verifying AI-generated outputs and delivering clear data stories to stakeholders. 1. Charts for Showing Change Over Time (Temporal Trends) These vis...
Whether you are stepping into your first role or preparing for a senior transition, technical interview loops for data roles demand a mix of statistical rigor, business intuition, and tool fluency. This guide outlines the top concepts, processes, and interview questions you will face, updated for today's modern data landscape. Part 1: Core Concepts & Statistical Foundations 1. What is an outlier, and how do you handle it? An outlier is a data point that differs significantly from the overall pattern of a dataset. Outliers can stem from entry errors, measurement flaws, or genuine high-variance events. Univariate Outliers: Deviate extremely within a single feature (e.g., an age recorded as 150). Typically identified via Z-Scores ( $\vert{}Z\vert{} > 3$ ) or the Interquartile Range Rule ( $1.5 \times \text{IQR}$ ). Multivariate Outliers: Deviate when multiple variables are combined (e.g., a 12-year-old with a PhD). Detected using distance metrics like Mahalanobis Distance...