Alfred Shingai

Alfred Shingai

Data analyst & software developer in Harare, Zimbabwe. Half my work is analysis — cleaning, modeling, and visualizing data. The other half is building the software that delivers it: a statistics lab, an open-data API, a source-tracked database. Final-year BSc Applied Statistics @ UZ.

stack

Python · FastAPI · Streamlit · pandas · Power BI · Excel — and pytest, because dashboards deserve tests too.

builds

Software, not just notebooks — Streamlit apps, FastAPI services, geospatial pipelines, and AI tooling in Deno + TypeScript. Tested, versioned, documented.

now

4+ projects shipped — 4 of them live on the internet — seeking data analyst roles.

//WORK

Software I design, build, and run solo — live, documented, or open source.

2026 — now

Who Builds Africa · open database

A source-tracked database of who finances, designs, builds, and operates African infrastructure. Every material claim links to a source, a verbatim passage, and an evidence-strength label.

2026

SokoData · open data API

One clean, documented API for African market prices — 486 markets, 31 commodities, daily USD movers, CPI and FX. Provenance per record, one ETL per dataset.

2026

Crisis Baseline Kit · geospatial toolkit

Pre-crisis baseline maps and databases from public data — type a country code, get boundaries, population, and zonal stats per district, with full provenance and quality gates that fail bad builds. 50+ African countries, rebuilt weekly by CI.

2026

Fundo · AI learning platform

100% free and open source, self-hostable: turn PDFs, DOCX, and markdown into study notes, flashcards, and quizzes — SM-2 spaced repetition, course mode, and any AI provider you point it at, or fully offline via Ollama. Deno + TypeScript.

2026

mini-polaris · LLM evals

Three realistic software-engineering eval tasks mined from my own shipped bugs — a StatLab header-only crash, SokoData ZiG false spikes, a WBA source-less claim. Each with starter code, hidden fixtures, and a 0–100 rubric graded deterministically in CI.

2025

StatLab Zim · statistics lab

An interactive statistics lab: upload a messy CSV, get a 13-metric summary, 6 charts, and 5 tests with plain-language decisions. No SPSS, no syntax. 19 pytest tests.

2023 — 2024

Data Analyst Intern · two placements

Sole analyst on retail and finance data — cleaned 10k+ rows, modeled in Power Query + DAX, shipped dashboards that replaced weekly manual reporting.

…and everything else anonymous AI-writing checks, terminal tools, frontend practice — 20+ public repositories. →

//CASE STUDIES
//THE JOB
▦ maize_prices_raw.csv — 6 rows
4 duplicates · 3 blanks · 2 formats
marketcommoditypricedatestatus
raw export → press cleaned.
//WRITING

Longer essays live on Substack — Learning as I Go: building data products, statistics without the jargon, and shipping small. Subscribe ↗

//STORY

I am finishing a BSc in Applied Statistics at the University of Zimbabwe. Two analyst internships taught me the job: clean 10k+ rows, keep the numbers honest, ship dashboards people actually open.

The software started because I kept wanting tools that didn't exist, so I taught myself to build them. Four products live on the internet now — a statistics lab, an open-data API for African markets, a geospatial crisis toolkit, a learning platform — and 4+ more shipped as open source: an LLM-eval harness, an anonymous AI-writing checker, a terminal timer. Twenty-plus repositories, tests before features.

The plan doesn't change after graduation: find the honest answer, then build the thing that delivers it.

//CONTACT

Open to interesting problems, collaborations, and conversations about data, dashboards, or the next StatLab feature. Best way in: a@alfredshingai.dev or tinonetsanaa@gmail.com — available remote or in-office.