Pick your capstone dataset
Use public data that is not from the course. Good places to look: government open-data sites, your city's open-data portal, and Kaggle datasets. Pick a question you care about first, then find data that answers it.
Milestone map
Milestone map
3 milestones
Complete Certificate Foundations and Demonstrate Spreadsheet Competence
4–6 weeks
The Google Data Analytics Certificate covers data foundations, spreadsheets, SQL, R programming, and data visualisation across eight courses (free audit via Coursera). Complete the first three courses. Then demonstrate practical spreadsheet competence by analysing a real public dataset — not the certificate's teaching datasets.
Proof required
Submit: (a) completion badges or screenshots from the first three certificate courses, and (b) a spreadsheet analysis on a real public dataset demonstrating: data cleaning (at least 3 data quality issues identified and resolved), pivot tables summarising by at least 2 dimensions, at least 2 charts with axis labels and interpretive titles, and a 300-word written summary of what the analysis revealed.
What gets checked
- Dataset is real and publicly available — not a certificate teaching dataset
- Data cleaning is documented — what issues were found and what action was taken
- Charts have interpretive titles that state the finding, not just describe the data axis
Common mistakes
- Using the certificate's own teaching datasets — the point is applying skills to a new independent dataset
- Charts with generic titles like 'Sales by Month' rather than interpretive titles like 'Q3 Sales Declined 18% Across All Categories'
Resources
What a verifier looks for
- Verify the independent analysis uses a different dataset from the certificate teaching materials — this is the key test of whether skills were actually learned.
- Review the data cleaning documentation: does it identify specific issue types and state what action was taken for each?
You'll sign in first, then come straight back here.
Complete SQL and R Modules with Independent Analyses
6–8 weeks
Complete the remaining certificate courses covering SQL (courses 4–5) and R programming (course 7). Build one independent SQL analysis on a publicly available database and one R analysis on a dataset of your choice — neither should use the certificate's teaching datasets.
Proof required
Submit: (a) completion screenshots for the SQL and R certificate courses, (b) a SQL script on a real public database (BigQuery Public Datasets, SQLite, or PostgreSQL) performing at minimum a JOIN across two tables, an aggregate with GROUP BY, and a filter — with a 200-word explanation of what it answered, and (c) an R script on a real dataset that reads data, cleans it, produces two ggplot2 visualisations, with a 300-word analysis of results.
What gets checked
- SQL script uses a real public database, not a local teaching database
- R script is reproducible — another analyst can run it on the same dataset
- Both written analyses state a finding — not just a description of what was plotted
Common mistakes
- SQL script uses SELECT * without filtering or aggregation
- R visualisations are default output of example code without modification
Resources
What a verifier looks for
- Review the R script for reproducibility: are all library() calls present? Does the data source reference a URL or public dataset rather than a local file?
- Check the SQL script: does it answer a real question, or is it a demonstration of syntax?
You'll sign in first, then come straight back here.
Complete the Certificate and Produce a Portfolio Case Study
6–8 weeks
Complete all eight courses of the Google Data Analytics Certificate including the capstone. Then produce an independent portfolio case study (distinct from the capstone) on a real-world dataset, following the full analysis lifecycle from question to recommendation. Present to a data or analytics professional who can challenge your methodology.
Proof required
Submit: (a) your Google Data Analytics Certificate completion credential, (b) an independent case study (minimum 1,000 words) presenting a business question, data source, cleaning and analysis methodology, visualisations, and a specific written recommendation, and (c) a Q&A log from presenting to a data professional (minimum 300 words with at least two methodological challenges and your responses).
What gets checked
- Certificate completion credential is from Google or Coursera — not a third-party mirror
- Case study is independent of the certificate capstone — new question, new dataset
- Recommendation is actionable and specific — 'further research is needed' is not a recommendation
Common mistakes
- Case study uses the same dataset as the certificate capstone
- Recommendation not grounded in the data
Powstik Guide
Your certificate and an independent portfolio case study.
Steps
- Finish every course and claim the credential.
- Pick a public dataset not used in the course.
- Run ask, prepare, process, analyse, share.
- Publish the case study.
Template
Credential link: Dataset (source): Question: Findings: Case study link:
What gets sent back
- Course dataset reused.
- No credential.
- Case study not public.
Resources
Foundationstart here
Depthgo deeper
What a verifier looks for
- Review the case study for the full data analysis lifecycle: question → data → cleaning → analysis → visualisation → recommendation. Missing steps suggest incomplete work.
- Your challenges should probe methodological choices: 'Why did you exclude these outliers?' 'How did you handle missing data in column X?'
You'll sign in first, then come straight back here.