airqoair 0.1.1

Openair-inspired Python tools for air quality analysis, visualisation, and mapping

airqoair

airqoair is a Python library for air-quality analysis, plotting, reporting, and interactive maps.

It is designed for environmental monitoring workflows where you want:

• direct input from pandas.DataFrame, .csv, .json, .jsonl, or .ndjson
• temporal diagnostics such as diurnal, weekday, monthly, and grouped site comparisons
• directional analysis such as wind roses, pollution roses, and polar plots
• interactive map outputs with folium
• no machine learning or forecasting in the package itself

Install

From PyPI:

pip install airqoair

Documentation:

https://airqoair-project.github.io/book

From this monorepo for development:

cd packages/airqoair
pip install -e ".[dev]"

Quick Start

import airqoair as aq

daily = aq.time_average(
    "kampala.csv",
    avg="D",
    pollutant_cols=["pm2_5", "pm10"],
)

variation = aq.time_variation("kampala.csv", pollutant="pm2_5")
monthly = aq.monthly_plot("kampala.csv", pollutant="pm2_5")
rose = aq.wind_rose("kampala.csv")
site_map = aq.station_map("kampala.csv", value="pm2_5")

print(daily.head())

variation.save("outputs/time_variation.png")
monthly.save("outputs/monthly_profile.png")
rose.save("outputs/wind_rose.png")
site_map.save("outputs/stations.html")

Grouped Conditioning

One of the most useful analysis patterns is conditioning the same workflow by site, year, season, or other categories.

airqoair supports this in the temporal workflow through:

• group_by="site_name" for an existing column in your data
• type="year", type="monthyear", type="season", type="weekday", or type="weekend" for derived date-based conditioning

Monthly bars by site:

import airqoair as aq

monthly_by_site = aq.monthly_plot(
    "kampala.csv",
    pollutant="pm2_5",
    group_by="site_name",
)

monthly_by_site.save("outputs/monthly_by_site.png")

Diurnal profile by season:

import airqoair as aq

diurnal_by_season = aq.diurnal_plot(
    "kampala.csv",
    pollutant="pm2_5",
    type="season",
)

diurnal_by_season.save("outputs/diurnal_by_season.png")

Monthly profile table by site:

import airqoair as aq

monthly = aq.monthly_profile(
    "kampala.csv",
    pollutant="pm2_5",
    group_by="site_name",
)

print(monthly.head())

Accepted Inputs

Most functions accept either a DataFrame or a file path.

Supported file formats:

• .csv
• .json
• .jsonl
• .ndjson

Examples:

import airqoair as aq

df = aq.load_data("kampala.jsonl")
daily = aq.time_average(df, avg="D", pollutant_cols=["pm2_5"])

import airqoair as aq

variation = aq.time_variation("kampala.json", pollutant="pm2_5")

Expected Columns

Defaults assume these names:

Purpose	Default column(s)
datetime	date
pollutant examples	pm2_5, pm10, no2, o3
wind speed	ws
wind direction	wd
latitude	latitude
longitude	longitude

You can override them in function calls, for example:

import airqoair as aq

aq.wind_rose("kampala.csv", ws="wind_speed", wd="wind_direction")
aq.station_map("kampala.csv", latitude="lat", longitude="lon", value="pm2_5")
aq.time_average("kampala.csv", date_col="timestamp", pollutant_cols=["pm2_5"])

Feature Coverage

Directional Analysis

• wind_rose()
• pollution_rose()
• percentile_rose()
• polar_frequency()
• polar_plot()
• polar_annulus()

Time Series, Profiles, and Reports

• time_average()
• time_variation()
• diurnal_profile()
• weekday_profile()
• monthly_profile()
• diurnal_plot()
• weekday_plot()
• monthly_plot()
• variation_report()

Visualization and Trends

• time_series_plot()
• distribution_plot()
• calendar_plot()
• time_proportion_plot()
• trend_level()
• run_regression()
• theil_sen_trend()
• smooth_trend()
• dilution_plot()
• pollutant_ratio_plot()
• trend_heatmap()

Model Evaluation

• model_evaluation()
• taylor_diagram()
• conditional_quantiles()

Interactive Maps

• station_map()
• map_primer()
• network_visualization()
• directional_map()
• polar_map()
• trajectory_analysis()

Common Workflows

Diurnal, Weekday, and Monthly Summaries

import airqoair as aq

diurnal = aq.diurnal_profile("kampala.csv", pollutant="pm2_5")
weekday = aq.weekday_profile("kampala.csv", pollutant="pm2_5")
monthly = aq.monthly_profile("kampala.csv", pollutant="pm2_5")

Grouped Site Comparison

import airqoair as aq

variation = aq.time_variation(
    "kampala.csv",
    pollutant="pm2_5",
    group_by="site_name",
)

report = aq.variation_report(
    "kampala.csv",
    pollutant="pm2_5",
    group_by="site_name",
)

print(report.sections["summary"].head())
variation.save("outputs/site_time_variation.png")

Directional Analysis

import airqoair as aq

aq.wind_rose("kampala.csv").save("outputs/wind_rose.png")
aq.percentile_rose("kampala.csv", pollutant="pm2_5").save("outputs/percentile_rose.png")
aq.polar_plot("kampala.csv", pollutant="pm2_5").save("outputs/polar_plot.png")

Time Trends

import airqoair as aq

aq.calendar_plot("kampala.csv", pollutant="pm2_5").save("outputs/calendar.png")
aq.smooth_trend("kampala.csv", pollutant="pm2_5", deseason=True, n_boot=200).save("outputs/smooth_trend.png")
aq.trend_heatmap("kampala.csv", pollutant="pm2_5").save("outputs/trend_heatmap.png")

Calendar plot with daily value labels and discrete bands:

import airqoair as aq

aq.calendar_plot(
    "kampala.csv",
    pollutant="pm2_5",
    year=2025,
    annotate="value",
    breaks=[0, 15, 35, 55, 150],
    labels=["Good", "Moderate", "Unhealthy", "Very Unhealthy"],
    lim=35,
).save("outputs/calendar_2025.png")

Time proportion plot split by site or wind sector over time:

import airqoair as aq

aq.time_proportion_plot(
    "kampala.csv",
    pollutant="pm2_5",
    proportion="site_name",
    avg_time="7D",
).save("outputs/time_proportion_by_site.png")

aq.time_proportion_plot(
    "kampala.csv",
    pollutant="pm2_5",
    proportion="wd",
    avg_time="1M",
    sector_count=8,
).save("outputs/time_proportion_by_wind_sector.png")

Trend-level heat map across month and hour:

import airqoair as aq

aq.trend_level(
    "kampala.csv",
    pollutant="pm2_5",
    x="month",
    y="hour",
).save("outputs/trend_level_month_hour.png")

Rolling regression diagnostics:

import airqoair as aq

aq.run_regression(
    "kampala.csv",
    x="ws",
    y="pm2_5",
    avg_time="MS",
    window=12,
).save("outputs/run_regression.png")

Theil-Sen trend with deseasonalising and confidence intervals:

import airqoair as aq

aq.theil_sen_trend(
    "kampala.csv",
    pollutant="pm2_5",
    deseason=True,
    n_boot=200,
).save("outputs/theil_sen_trend.png")

Smooth trend with deseasonalising and grouped site panels:

import airqoair as aq

aq.smooth_trend(
    "kampala.csv",
    pollutant="pm2_5",
    group_by="site_name",
    deseason=True,
    n_boot=200,
).save("outputs/smooth_trend_by_site.png")

Maps

import airqoair as aq

aq.station_map("kampala.csv", value="pm2_5").save("outputs/stations.html")
aq.directional_map("kampala.csv").save("outputs/directional_map.html")
aq.polar_map("kampala.csv", site_col="site_name", pollutant="pm2_5", kind="polar_plot").save("outputs/polar_map.html")

Grouped interactive map with clustering, marker scaling, and a heat layer:

import airqoair as aq

aq.station_map(
    "kampala.csv",
    value="pm2_5",
    group_by="site_name",
    radius_col="device_count",
    cluster=True,
    heatmap=True,
    tiles=["CartoDB positron", "OpenStreetMap"],
).save("outputs/stations_grouped.html")

Returned Objects

Plotting functions return AirQoFigure.

• .data contains the processed table behind the plot
• .figure is the underlying Matplotlib figure
• .ax returns the first Matplotlib axis for single-axis convenience
• .axes returns all Matplotlib axes on the figure
• .show() displays the plot
• .save(path) writes the plot to disk

Map functions return AirQoMap.

• .data contains the mapped rows
• .map is the underlying folium.Map
• .save(path) writes the HTML map

variation_report() returns AirQoReport.

• .sections contains summary and profile tables
• .save_markdown(path) writes a markdown report
• .save_figure(path) writes the report figure

Current Scope

airqoair is designed for practical air-quality diagnostics in Python, but it is not yet a full feature-for-feature replacement for every established toolkit in this space.

This package currently focuses on:

• practical temporal profiles and grouped conditioning
• directional analysis and mapping
• trend and evaluation utilities that are straightforward to use from Python data workflows

Development

Run tests from the package root:

cd packages/airqoair
python -m pytest tests

If you are working from the repo without installation, point PYTHONPATH at packages/airqoair, not the repository root.

Documentation site

A book-style documentation site scaffold is included under:

packages/airqoair/docs/
packages/airqoair/mkdocs.yml

To build it locally:

cd packages/airqoair
pip install -e ".[docs]"
python -m mkdocs serve

Images for the docs can be stored in:

packages/airqoair/docs/assets/images/

.gitignore

The package-local .gitignore excludes:

• Python caches and bytecode
• virtual environments
• pytest, coverage, mypy, ruff, tox, and nox caches
• build artifacts and editable-install metadata
• generated plots, HTML maps, markdown reports, and temporary outputs

License

MIT