Turn raw data into business decisions.
3 topics
Excel remains the most widely used data analysis tool in business. Master: formulas (VLOOKUP, INDEX/MATCH, SUMIFS, COUNTIFS), pivot tables for summarizing large datasets, conditional formatting for visual patterns, data validation for clean input, and charts (bar, line, scatter, combo). Learn keyboard shortcuts to work 3x faster. Google Sheets adds collaboration and Apps Script for automation. Power Query (Get & Transform) handles data cleaning — merging files, unpivoting columns, removing duplicates, and transforming data types. For large datasets, understand Excel's row limits (1M rows) and when to graduate to SQL or Python. Excel is your prototyping tool — build the analysis in Excel first, then automate with code.
2 resources
SQL is the single most important skill for data analysts. You need to query relational databases fluently. Master: SELECT with WHERE, GROUP BY with aggregate functions (COUNT, SUM, AVG, MIN, MAX), JOIN types (INNER, LEFT, FULL), subqueries and CTEs (Common Table Expressions) for readability, window functions (ROW_NUMBER, RANK, LAG, LEAD, running totals, moving averages), CASE statements for conditional logic, date functions for time-based analysis, and HAVING for filtering aggregated results. Learn to write performant queries — understand indexes, EXPLAIN plans, and avoid SELECT *. Practice with real datasets on platforms like Mode Analytics, SQLZoo, or LeetCode SQL. Most companies use PostgreSQL, MySQL, BigQuery, Snowflake, or Redshift.
Statistics is the backbone of data analysis — it helps you draw valid conclusions from data. Essential concepts: descriptive statistics (mean, median, mode, standard deviation, percentiles, distributions), probability (conditional probability, Bayes theorem), hypothesis testing (null and alternative hypotheses, p-values, confidence intervals, t-tests, chi-square tests), correlation vs causation (just because two things are related does not mean one causes the other), regression analysis (linear regression for predicting continuous outcomes), and A/B testing (comparing two variants to determine which performs better with statistical significance). Understand sampling bias, survivorship bias, and Simpson's paradox — these trap even experienced analysts. Statistics helps you answer "is this result real or just noise?"
2 topics
Python is the analyst's Swiss Army knife. Pandas is the core library for data manipulation — DataFrames let you load, clean, transform, and analyze tabular data. Key operations: reading CSV/Excel/JSON files, filtering rows, selecting columns, grouping and aggregating, merging/joining datasets, handling missing values (fillna, dropna), string operations, datetime parsing, and applying custom functions. NumPy provides the numerical foundation — arrays, vectorized operations, linear algebra, and random number generation. Learn method chaining for readable Pandas code: df.query(...).groupby(...).agg(...).sort_values(...). Understand the difference between copies and views to avoid SettingWithCopyWarning. Jupyter Notebooks provide an interactive environment for exploratory analysis.
Visualization transforms numbers into insights that stakeholders can understand and act on. Matplotlib is the foundational Python plotting library — highly customizable but verbose. Seaborn builds on Matplotlib with beautiful statistical visualizations (heatmaps, pair plots, violin plots, box plots). Plotly creates interactive charts (hover tooltips, zoom, filter) that work great in dashboards and presentations. Know which chart to use: bar charts for comparisons, line charts for trends over time, scatter plots for relationships, histograms for distributions, heatmaps for correlations, and pie charts (sparingly) for proportions. Follow visualization best practices: clear titles, labeled axes, appropriate color schemes, no 3D effects, and highlight the key insight. Edward Tufte's principles — maximize data-to-ink ratio — should guide your design.
Business Intelligence tools let you create interactive dashboards without code. Tableau is the industry leader for visual analytics — drag-and-drop interface, calculated fields, parameters, and stunning visualizations. Power BI integrates deeply with Microsoft ecosystem (Excel, SharePoint, Azure) and offers DAX (Data Analysis Expressions) for complex calculations. Looker (Google Cloud) uses LookML to define a semantic data model, ensuring consistent metrics across the organization. Learn: connecting to data sources, building calculated fields, creating interactive filters, designing dashboard layouts, publishing and sharing reports, and setting up scheduled refreshes. The key skill is translating business questions into visual answers — start with the question, not the chart.
The most important analyst skill is not technical — it is communicating findings in a way that drives action. Data storytelling combines data, visuals, and narrative. Structure your analysis as: context (what is the business question?), findings (what does the data show?), implications (what does it mean?), and recommendations (what should we do?). Use the pyramid principle — lead with the conclusion, then supporting evidence. Design presentations for your audience: executives want key metrics and decisions, managers want trends and exceptions, and technical teams want methodology and detail. Avoid data dumps — curate the 3-5 most important insights. Practice presenting uncertainty honestly — confidence intervals, caveats, and what the data does not show.