openmarina 0.0.1

The open marina for maritime data — one clean Python interface over fragmented US public marine data (NDBC, tides & currents, and more).

openmarina

Clean marine data in one line — no parsers, no unit conversions, no agency quirks.

import openmarina

cf = openmarina.load("ndbc:41122")     # a buoy off Florida, last 24h
df = cf.to_wide()                       # one clean column per variable

---

The problem it solves

You want recent wave height and wind near a spot on the water. Simple — until you try.

One agency hands you whitespace-delimited text with a cryptic header. The next wants an API key and returns JSON. Pressure is in hectopascals here, millibars there. Wind direction means "coming from" in one feed and "going to" in another. Timestamps arrive in three different time zones. So you write a parser — and patch it again next month when the format quietly changes. Then you need a second agency, and you start over.

Public ocean data is free, but using it costs you days.

openmarina does that work once, so you don't. One call returns a clean, normalized table — the same variable names, the same SI units, UTC, WGS84 — no matter which agency the data came from.

How it works

• Normalized once. A wave height from a Florida buoy and one from a Maine buoy come back as the same variable, in the same units, in the same table.
• Quality you can see. Every value carries a qc_flag (good / suspect / missing), so you know what to trust.
• Find what's near you. nearest(lat, lon) — or nearest --zip 33139.
• Add an agency in one PR. Each source is a small adapter behind one shared contract, verified by a built-in conformance harness. New sources don't complicate the ones you already use.

Get started

Install (Python 3.10+):

pip install openmarina        # once published
pip install -e ".[dev]"       # from a clone

From Python:

import openmarina

cf = openmarina.load("coops:8723214", start="2026-06-23", end="2026-06-24")
cf.to_wide()                                        # tidy table, one column per variable
cf.meta                                             # provenance, datum, cadence
openmarina.load_many(["ndbc:41122", "ndbc:41009"])  # several stations at once
openmarina.nearest_zip("33139")                     # nearest station to a ZIP (pip install pgeocode)

From the terminal:

openmarina pull ndbc:41122 --wide -o out.csv
openmarina nearest 25.76 -80.19
openmarina stations
# on a locked-down machine, run any command as:  python -m openmarina ...

Sources today

Source	What it gives you
NDBC	NOAA buoys — waves, wind, water/air temperature, pressure
CO-OPS	NOAA Tides & Currents — water level, currents, met
ERDDAP	one configurable adapter for 60+ global providers

New agencies are welcome — the 15-minute path is in CONTRIBUTING.md.

License

Apache-2.0 — free to use, including commercially.