Data analyst · Statistics · Storytelling

Turning messy data into clear decisions.

I’m Annisa, an aspiring data analyst combining statistical thinking, Python, R and digital product experience to uncover patterns—and explain what they mean to the people making decisions.

Explore my work Download résumé

Building a cross-industry analytics portfolio

PORTFOLIO_SIGNAL_01

Housing records34,857cleaned and explored

Current focus4industry case studies

From data to decisionsignal / noise

Tools I use and develop

Python

SQL

SPSS

MATLAB

Selected work

Projects built around decisions, not just charts.

Each case study is designed to show the full analytical journey: the question, data quality, method, findings, limitations and a recommendation someone could act on.

01 · Real estate analytics

A reproducible exploration of Melbourne property records, suburb-level patterns and the factors associated with sale price.

Current flagship · being refined

Where does Melbourne offer value?

The project begins with a practical question: how can a buyer compare suburbs after considering property characteristics and location—not simply rank the highest prices?

34,857raw records

21source variables

Pythonanalysis stack

Data cleaningEDARegression plannedGeospatial planned

Open notebook ↗ View Canva deck ↗

Read the upgrade brief

Make data loading reproducible and document the source.
Model log price while checking assumptions and validation error.
Avoid treating price-derived ratios as independent causes of price.
Finish with stakeholder recommendations and explicit limitations.

02 · Retail & e-commerce

Which customers create durable value?

A customer analytics case study using SQL, RFM segmentation and cohort retention to support targeted engagement decisions.

In development

SQLRFMDashboard

03 · Operations

Where will demand exceed available stock?

A forecasting project comparing simple and seasonal baselines through rolling backtests, then translating error into inventory risk.

In development

Time seriesBacktestingPython

04 · Product experimentation

Did the new experience improve conversion?

An end-to-end experiment covering power, sample size, effect estimation, uncertainty and segment interactions—not only a p-value.

In development

A/B testingANOVADecision memo

How I work

Analysis that remains understandable after the meeting.

My goal is to connect sound statistical reasoning with an explanation that a non-technical stakeholder can use confidently.

01 / ANALYSE

Find the signal

Structure questions, screen data, quantify uncertainty and choose a method that matches the design rather than forcing the data into a fashionable model.

PythonRSQLRegressionANOVA

02 / EXPLAIN

Tell the truth clearly

Turn output into direct findings, separate association from causation, make limitations visible and use visual hierarchy to keep attention on the decision.

Data visualisationCanvaExecutive summaries

03 / BUILD

Make insight usable

Combine analytics with prior web-development experience to create reproducible notebooks, interactive prototypes and accessible portfolio experiences.

GitHubColabHTML/CSSStreamlit planned

About Annisa

A statistical mind with a builder’s background.

I’m pursuing a Master of Science in Mathematics and Statistics, strengthening my work in statistical modelling, experimental design, multivariate analysis and data visualisation.

Before focusing on analytics, I built digital products and websites. That experience still shapes how I work: I care about whether an analysis is reproducible, whether the result is communicated clearly and whether the final output is genuinely usable.

Current studyMathematics and Statistics

Analytical focusEvidence-based decisions

Previous foundationWeb and product development

ValuesClarity, curiosity and integrity

LinkedIn ↗ GitHub ↗

A repeatable process

From question to recommendation.

The same structure is used across every portfolio case study, making the reasoning easy to review and the projects easy to compare.

STEP 01

Frame

Define the stakeholder, decision and measurable question.

STEP 02

Validate

Understand provenance, missingness, bias and measurement limitations.

STEP 03

Analyse

Use an appropriate method, diagnostics and honest comparisons.

STEP 04

Recommend

Translate findings, uncertainty and limitations into the next action.

Start a conversation

Have a data question worth untangling?

I’m interested in analyst opportunities, portfolio collaborations and thoughtful conversations about statistics, products and how evidence supports better decisions.

Email me ↗