Artificial Intelligence & Data Science · 1.1

Data Analyst

Turn raw data into actionable business insights through querying, visualization, and statistical analysis.

5–6 months6 phases31 skills

Foundations

Mathematics refresherdescriptive statistics (mean, median, mode, standard deviation, variance, percentiles, IQR), probability basics (Bayes' theorem, conditional probability), distributions (normal distribution and the 68-95-99.7 rule, binomial distribution for conversion-rate modeling, Poisson distribution for event-rate modeling)

Introduction to data types and their implications

Introduction to data types and their implicationscategorical (nominal vs ordinal), numerical (discrete vs continuous), time-series — why data type determines which statistical tests and visualizations are valid

Microsoft Excel mastery

Microsoft Excel masteryVLOOKUP, INDEX-MATCH, XLOOKUP, pivot tables with calculated fields and slicers, conditional formatting with data bars and icon sets, data validation with dropdown lists, SUMIFS/COUNTIFS for multi-criteria aggregation, named ranges, sparklines, and combo charts

Google Sheets proficiency

Google Sheets proficiencyQUERY function (SQL-like filtering and aggregation inside a spreadsheet), IMPORTRANGE for cross-spreadsheet data, ARRAYFORMULA for column-wide calculations, Google Apps Script basics for automating recurring reports

Business fundamentals

Business fundamentalsKPIs vs vanity metrics, funnel analysis (AARRR framework), cohort analysis (grouping users by signup month and tracking retention), revenue metrics (MRR, ARR, churn rate, CLV, CAC), unit economics

Phase 02 · 17h

SQL & Databases

Relational database concepts

Relational database conceptstables, schemas, normalization (1NF through 3NF with examples), primary and foreign keys, ER diagrams, when denormalization is acceptable for analytics

SQL fundamentals in PostgreSQL

SQL fundamentals in PostgreSQLSELECT, WHERE with comparison and logical operators, GROUP BY, HAVING, ORDER BY, LIMIT/OFFSET, DISTINCT, aggregate functions (COUNT, SUM, AVG, MIN, MAX)

Joins in depth

Joins in depthINNER JOIN, LEFT JOIN (with NULL detection for finding non-matching rows), RIGHT JOIN, FULL OUTER JOIN, CROSS JOIN (generating date×product matrices), self-joins (hierarchical data like org charts)

Advanced SQL

Advanced SQLsubqueries (scalar, column, table), CTEs using WITH for readable multi-step queries, window functions (ROW_NUMBER, RANK, DENSE_RANK, LAG, LEAD for month-over-month growth, SUM OVER for running totals, AVG OVER with ROWS BETWEEN for moving averages), CASE WHEN for conditional logic inside queries

Google BigQuery as the cloud warehouse

Google BigQuery as the cloud warehousesetting up access (1 TB queries/month), querying public datasets (google_analytics_sample, stackoverflow), cost optimization (SELECT only needed columns, use partitioned tables, avoid SELECT *)

Practice platforms

Practice platformscomplete 50+ problems on LeetCode SQL (Medium/Hard), 30 on HackerRank SQL, 20 on StrataScratch (real company interview questions)

Phase 03 · 11h

Python for Analysis

Python environment setup

Python environment setupinstall Python 3.12 via python.org, use VS Code with Python and Jupyter extensions, create virtual environments with python -m venv, install packages with pip, use Google Colab for zero-setup notebooks

Python basics

Python basicsvariables, data types, f-strings, collections (lists, dictionaries, tuples, sets), control flow, functions with default parameters, list comprehensions, error handling with try/except

Pandas deep-dive

Pandas deep-diveDataFrames and Series, reading CSV/JSON/Excel/Parquet files, filtering with boolean indexing and .query(), sorting, column operations with .apply() and lambda, GroupBy with .agg() for multiple aggregations, merging DataFrames (equivalent to SQL joins), pivot tables, reshaping with .melt() and .pivot(), date operations with .dt accessor and pd.Grouper

Data cleaning

Data cleaningdetecting and handling missing values (.isnull().sum(), .dropna(), .fillna() with median/mode), removing duplicates (.drop_duplicates()), outlier detection with IQR method, type casting with .astype(), string cleaning with .str accessor methods

NumPy fundamentals

NumPy fundamentalsarray creation and operations, vectorized arithmetic (10-100x faster than Python loops), broadcasting, statistical functions (np.mean, np.std, np.percentile, np.corrcoef), random number generation for simulations

File formats

File formatsCSV (human-readable but slow), JSON (nested structures from APIs, flatten with pd.json_normalize()), Parquet (columnar binary format — 5-10x smaller than CSV, preserves types, much faster reads)

Phase 04 · 17h

Data Visualization

Visualization principles

Visualization principlesTufte's data-ink ratio (remove chartjunk), chart selection framework (line for trends, bar for comparisons, scatter for relationships, histogram for distributions, heatmap for correlations), ColorBrewer palettes for accessibility, designing for colorblind readers

Matplotlib and Seaborn

Matplotlib and Seabornuse the object-oriented interface (fig, ax = plt.subplots()), line/bar/scatter/histogram/box plots, subplots for multi-chart layouts, customizing titles/labels/legends/annotations, saving at 300 DPI; Seaborn for statistical defaults — sns.histplot with KDE overlay, sns.heatmap for correlation matrices with annotations, sns.boxplot/violinplot for distribution comparisons, sns.pairplot for exploring all variable relationships

Plotly for interactive visualizations

Plotly for interactive visualizationsPlotly Express for quick interactive charts (px.scatter, px.line, px.bar with hover_data for details), px.choropleth for geographic data, building a simple Dash dashboard with dropdown filters and callback-driven chart updates

Tableau Desktop (using Tableau Public)

Tableau Desktop (using Tableau Public)connecting to CSV, Excel, and PostgreSQL data sources, understanding dimensions vs measures, building bar/line/scatter/map/treemap charts, creating calculated fields, Level of Detail (LOD) expressions (FIXED for calculating at a specific granularity regardless of view, INCLUDE and EXCLUDE for adding/removing dimensions), table calculations (running total, percent of total, moving average), parameters for user-driven metric switching, building multi-sheet dashboards with cross-filters and dashboard actions, publishing to Tableau Public

Microsoft Power BI Desktop

Microsoft Power BI Desktopimporting data from Excel/CSV/PostgreSQL, Power Query Editor for ETL (rename columns, change types, merge/append queries, unpivot), star schema data modeling with relationships, DAX measures (SUM, CALCULATE for modifying filter context, FILTER, ALL, time intelligence with TOTALYTD and SAMEPERIODLASTYEAR for YoY growth), building multi-page reports with slicers, cards, KPI visuals, matrix, decomposition tree, drillthrough pages, publishing to Power BI Service

Storytelling with data

Storytelling with datastructuring presentations using situation-complication-resolution, choosing charts for the audience (executives want trends and KPIs, analysts want granular breakdowns), annotation techniques (highlighting key insights with callout text on charts)

Phase 05 · 8h

Statistics & Analysis Techniques

Inferential statistics

Inferential statisticshypothesis testing framework (null vs alternative hypothesis, significance level α=0.05, p-values, Type I and Type II errors), t-tests (one-sample, two-sample independent, paired), chi-square test for independence between categorical variables, ANOVA for comparing means across 3+ groups — all implemented with scipy.stats in Python

Correlation vs causation

Correlation vs causationPearson and Spearman correlation coefficients, confounding variables, Simpson's Paradox (how aggregate trends can reverse in subgroups) with a worked real-world example

A/B testing

A/B testingdefining the metric and minimum detectable effect, calculating sample size with power analysis (statsmodels.stats.power), proper randomization, running the test for full-week cycles, analyzing with two-proportion z-test, reporting confidence intervals, avoiding common pitfalls (peeking, multiple comparisons, novelty effects)

Regression basics

Regression basicssimple and multiple linear regression with sklearn and statsmodels (interpreting coefficients, R-squared, adjusted R-squared, p-values per coefficient), checking assumptions with residual plots

Time-series fundamentals

Time-series fundamentalsdecomposing into trend, seasonality, and residual (statsmodels.tsa.seasonal_decompose), moving averages (SMA and EMA), YoY and MoM growth in Pandas, basic forecasting with Prophet (installing, fitting, interpreting output)

Phase 06 · 13h

Capstone

Capstone Project 1

Capstone Project 1End-to-end e-commerce analysis using the Brazilian E-Commerce (Olist) Kaggle dataset — clean 9 merged CSVs, perform EDA, build RFM customer segmentation, create a 4-view Tableau dashboard, write a 2-page executive summary with recommendations

Capstone Project 2

Capstone Project 2A/B test analysis — validate randomization, calculate conversion rates, perform chi-square test, compute confidence intervals, assess sample adequacy, write a recommendation memo

Capstone Project 3

Capstone Project 3Power BI report for a fictional SaaS company — star schema data model (fact_subscriptions + dimension tables), DAX measures for MRR, churn rate, net revenue retention, 3-page report with slicers and drillthrough

Ready to start Data Analyst?

Start this roadmap to track progress and generate courses for each skill. Login required only to begin.