Muhammad Gardian Novandri Portfolio
AML Reporting Specialist at Bank Jago, after five years as a loan and asset-management analyst at Bank BTN. A Canva-built portfolio proves that four well-packaged analyses — each closing on a number — beat a custom-coded site.
Location
Didn't say. We're respecting the mystery.
Skills
What makes it work
A breakdown of the choices that make this portfolio stand out.
A banking career paired with cross-industry projects
The day job is bank-side — AML reporting at Bank Jago after five years in loan and asset-management analysis at Bank BTN — but the four showcased projects deliberately leave banking: ride-hailing pricing, employee attrition, a fast-food A/B test, e-commerce retention. The career supplies regulated-industry credibility; the projects prove the same analytical toolkit transfers to any business question.
Quantified business impact with specificity
Each project closes on a specific figure: +11.5% revenue from the Uber pricing model, a predicted 3.41% drop in IBM attrition, and a 5.8%–23.4% expected retention lift from RFM treatments. These are not generic "improved efficiency" claims — the specificity is what separates a convincing portfolio from a credential list.
Full analyst stack represented coherently
SQL for data extraction, Python for analysis, Tableau and Google Looker for stakeholder-facing dashboards — the stack covers every layer of the analyst workflow from raw query to executive presentation. The combination signals that this analyst can own the full pipeline, not just one slice of it.
Canva as a deliberate tool choice
Using Canva instead of a custom-coded site is not a limitation — it is a sensible prioritisation. The portfolio looks polished and professional because the effort went into content and structure rather than infrastructure. Hiring managers evaluating a data analyst are evaluating the analysis, not the web framework.
What Data Analysts can take from this
Specific, actionable tips to apply to your own portfolio — no generic advice.
If your job history already signals one industry, use portfolio projects to show range beyond it — employers see the domain experience on the résumé and the transferable skills in the projects.
Attach a real business metric to every project — revenue, attrition, retention — using the exact KPI the stakeholder tracks, not generic "improvement" language.
Show the full stack. SQL → Python → Tableau/Looker mirrors the actual analyst workflow. Portfolios that only show dashboards leave employers wondering whether you can source and transform the data yourself.
Canva, Notion, or a well-structured PDF portfolio can outperform a custom-coded site if the content is stronger. Do not let portfolio infrastructure become a procrastination mechanism.
Name the industry of your day job clearly in your headline even when your projects are elsewhere — regulated-industry experience like banking or AML is a real differentiator, and it should not be left for the reader to infer.
Projects
Deck, dashboard, notebook. Muhammad Gardian Novandri packages each analysis in more than one format and closes most on a number rather than an adjective — the way findings have to travel when the audience will never open the code.
Uber Pricing Optimization in New York City
An end-to-end analysis to find the best pricing model for Uber bookings in New York City. Linear regression measures the impact of each booking variable. The existing pricing model (high-fare and discounted-fare applications) is then evaluated, an impact analysis is run, and a pricing recommendation follows.
Evaluating an existing pricing scheme before proposing a new one, and closing on a revenue number, is how pricing analysis is actually presented to stakeholders.
Closes on a single business number: +11.5% revenue
Three deliverables: deck, Tableau dashboard, Colab code
Evaluates the current model before recommending a new one
IBM Employee Attrition Prediction and Strategy 2023
Logistic regression predicts attrition among existing IBM employees. That prediction feeds a retention strategy aimed at fairer working conditions, and the analysis estimates how much the strategy would reduce attrition.
Going from a classifier to a costed strategy with an expected effect size is the step many modelling projects skip.
Model output turned into an HR strategy, not left as a probability
Expected impact quantified (−3.41% attrition)
Deck, dashboard and notebook formats
Fast Food Marketing Campaign A/B Testing
Determines the most profitable of three marketing campaigns run by a fast-food chain whose stores fall into three market-size categories. ANOVA is used as the hypothesis test, followed by post-hoc t-tests.
A proper multi-arm test (ANOVA, then post-hoc comparisons) instead of eyeballing averages, done entirely in spreadsheet tools most business teams already use.
Three-arm test with market size as a factor
ANOVA followed by post-hoc t-tests
Built entirely in Google Sheets/Excel
E-Commerce Cohort Analysis and RFM Segmentation
A project to boost an e-commerce company's retention rate. Customers are grouped with RFM segmentation, with a treatment strategy defined per segment. Cohort analysis then finds the maximum range of CLTV, CAC and CRC for the marketing team to use.
SQL in BigQuery through to Looker dashboards covers the full analyst stack, and the output is a per-segment action plan rather than a chart.
BigQuery SQL → Colab cleaning → Looker Studio dashboards
Retention strategy per RFM segment, with a stated expected range
Cohort-based CLTV/CAC/CRC ranges handed to marketing
More Data Analyst portfolios
Ready to build your portfolio?
Follow the Data Analyst roadmap — skills, projects, and timeline to get hired.