data-analytics

Workshop materials for “Ready, Set, Publish: Write Your First Python Package.” This package is published to TestPyPI and consumed by kolbl/streamlit-app, a Streamlit dashboard that uses the code from this repo.

---

Development setup

UV

Install uv (pinned installer example):

curl -LsSf https://astral.sh/uv/0.11.1/install.sh | sh -s

Windows installers

The installer puts uv under %USERPROFILE%\.local\bin. If new terminals still do not find uv, persist that directory on your User Path (PowerShell):

# Persist for your user account; run in PowerShell:
$uvBin = Join-Path $env:USERPROFILE ".local\bin"
[Environment]::SetEnvironmentVariable(
  "Path",
  "$uvBin;" + [Environment]::GetEnvironmentVariable("Path", "User"),
  "User"
)

Close all terminals, open a new one, then verify:

uv --version

Environment and dependencies

uv sync

A .venv is created. In your editor, select that interpreter (or activate the venv if you prefer).

Add packages:

uv add pandas
uv add pytest --group dev
uv sync --group dev

Tests

uv run pytest

In Cursor/VS Code: open Testing and use “Focus on Test Explorer View” (or search for “Focus on Test View”).

Pre-commit

pre-commit install
pre-commit run --all-files

If pre-commit is not on your PATH, use uv run pre-commit install and uv run pre-commit run --all-files.

---

Usage

Column max, min, and span

calculate_max_of_column(df, column) is implemented as df[column].max(): pandas returns the largest value with its default skipna=True, so missing values are ignored. If every cell in the column is missing, the result is NaN.

The same idea applies to calculate_min_of_column (.min()) and calculate_span_of_column, which is max − min on that column (again via .max() and .min(), so NaNs are skipped consistently).

import pandas as pd
from data_analytics.column_statistics import (
    calculate_max_of_column,
    calculate_min_of_column,
    calculate_span_of_column,
)

df = pd.DataFrame({"x": [1.0, 5.0, 3.0, float("nan")]})
calculate_max_of_column(df, "x")   # 5.0
calculate_min_of_column(df, "x")   # 1.0
calculate_span_of_column(df, "x")  # 4.0

---

Publishing (TestPyPI and GitHub)

TestPyPI account

• Register: test.pypi.org/account/register/
• In account settings, connect GitHub if you use that integration.

GitHub requirements

• Secret PYPI_API_TOKEN: PyPI → Account → API tokens (classic) with upload scope for this project, or a project-scoped token.
• Project on PyPI whose name matches what you upload (here the distribution is data-analytics / import name data_analytics).

If PYPI_API_TOKEN is missing, the publish step fails until you add it.

Versioning: Every push to main that passes CI may upload a new dev version (0.1.0.dev1, 0.1.0.dev2, …). PyPI does not allow reusing the same version.

---

GitHub Pages (MkDocs)

Enable Pages on the repository:

1. Repository settings: github.com/kolbl/data-analytics/settings/pages
2. Build and deployment → Source: choose GitHub Actions.

After Pages is enabled and permissions match your workflow, deployments should succeed.