OWPHydroTools :: WaterData Client

This subpackage implements an interface to the USGS WaterData APIs. The primary use for this tool is to populate pandas.Dataframe objects with USGS streamflow data. See the WaterData Client Documentation for a complete list and description of the currently available methods. To report bugs or request new features, submit an issue through the OWPHydroTools Issue Tracker on GitHub.

Installation

In accordance with the python community, we support and advise the usage of virtual environments in any workflow using python. In the following installation guide, we use python's built-in venv module to create a virtual environment in which the tool will be installed. Note this is just personal preference, any python virtual environment manager should work just fine (conda, pipenv, etc. ).

# Create and activate python environment, requires python >= 3.11
$ python3 -m venv venv
$ source venv/bin/activate
$ python3 -m pip install --upgrade pip

# Install waterdata_client
$ python3 -m pip install hydrotools.waterdata_client

Usage

The following examples demonstrate typical usage of hydrotools.waterdata_client.

Retrieve most recent streamflow values

# The LatestContinuousClient provides an interface to the latest-continuous endpoint.
from hydrotools.waterdata_client import LatestContinuousClient

# The default client returns deserialized JSON as a list of Python dictionaries.
#   We'll import a transformer that converts these data to a GeoDataFrame.
from hydrotools.waterdata_client.transformers import to_geodataframe

# Instantiate the client with the GeoDataFrame transformer
client = LatestContinuousClient(transformer=to_geodataframe)

# Retrieve data
observations = client.get(
    monitoring_location_id="USGS-02146470",
    parameter_code="00060" # Volumetric streamflow in ft^3/s
    )

# Look at values
print(observations[["usgs_site_code", "value_time", "value", "geometry"]])
print(observations.columns)

Output

  usgs_site_code                 value_time value                    geometry
0  USGS-02146470  2026-05-18T18:45:00+00:00  0.40  POINT (-80.85306 35.16444)

Index(['geometry', 'id', 'time_series_id', 'usgs_site_code', 'parameter_code',
       'statistic_id', 'value_time', 'value', 'measurement_unit',
       'approval_status', 'qualifiers', 'last_modified'],
      dtype='str')

Retrieve values using datetime objects

import pandas as pd
# from datetime import datetime, timedelta
# from zoneinfo import ZoneInfo

# The ContinuousClient provides an interface to the continuous endpoint.
from hydrotools.waterdata_client import ContinuousClient

# The default client returns deserialized JSON as a list of Python dictionaries.
#   We'll import a transformer that converts these data to a DataFrame.
from hydrotools.waterdata_client.transformers import to_dataframe

# Instantiate the client with the DataFrame transformer
client = ContinuousClient(transformer=to_dataframe)

# Retrieve data using datetime handling
observations = client.get(
    monitoring_location_id="USGS-02146470",
    limit=50_000,
    parameter_code="00065", # Stage in ft.

    # Strings are passed directly to the API without intervention. See the API
    #   documentation for more information about valid string formats (RFC 3339).
    # datetime="2025-05-03T12:00:00-05:00/2025-05-05T00:00:00-05:00",

    # Retrieve the last 6 hours using pandas.Timedelta
    datetime=pd.Timedelta(hours=6),

    # Retrieve a specific datetime using pandas.Timestamp
    #   Timezone naive objects are assumed to be UTC.
    # datetime=pd.Timestamp("2010-01-01"),

    # Timezone aware types are converted to UTC before retrieval.
    # datetime=pd.Timestamp("2010-01-01", tzinfo=ZoneInfo("America/Chicago")),

    # Retrieve using a standard datetime object
    # datetime=datetime(2026, 1, 1, 0, 30, 0),

    # Retrieve using a standard timedelta object
    # datetime=timedelta(hours=6),

    # Use a list of datetime objects to specify a bounded interval [start, end]
    # datetime=[pd.Timestamp("2026-01-01"), pd.Timestamp("2026-01-02")],

    # Specify a half-bounded interval using None
    # datetime=[pd.Timestamp("2026-05-01"), None],
    )

# Look at values
print(observations[["usgs_site_code", "value_time", "value", "geometry"]])
print(observations.columns)

Output

   usgs_site_code                 value_time value geometry
0   USGS-02146470  2026-05-18T13:00:00+00:00  3.07     None
1   USGS-02146470  2026-05-18T13:05:00+00:00  3.07     None
2   USGS-02146470  2026-05-18T13:10:00+00:00  3.07     None
3   USGS-02146470  2026-05-18T13:15:00+00:00  3.08     None
4   USGS-02146470  2026-05-18T13:20:00+00:00  3.08     None
..            ...                        ...   ...      ...
66  USGS-02146470  2026-05-18T18:30:00+00:00  3.08     None
67  USGS-02146470  2026-05-18T18:35:00+00:00  3.08     None
68  USGS-02146470  2026-05-18T18:40:00+00:00  3.08     None
69  USGS-02146470  2026-05-18T18:45:00+00:00  3.08     None
70  USGS-02146470  2026-05-18T18:50:00+00:00  3.08     None

[71 rows x 4 columns]
Index(['type', 'id', 'geometry', 'geo_feature_id', 'time_series_id',
       'usgs_site_code', 'parameter_code', 'statistic_id', 'value_time',
       'value', 'measurement_unit', 'approval_status', 'qualifiers',
       'last_modified'],
      dtype='str')