pyinaturalist

Introduction

iNaturalist is a community science platform that helps people get involved in the natural world by observing and identifying the living things around them. Collectively, the community produces a rich source of global biodiversity data that can be valuable to anyone from hobbyists to scientists.

pyinaturalist is a client for the iNaturalist API that makes these data easily accessible in the python programming language.

• Features
• Quickstart
• Next Steps
• Feedback
• Related Projects

Features

• ➡️ Easier requests: Simplified request formats, easy pagination, and complete request parameter type annotations for better IDE integration
• ⬅️ Convenient responses: Type conversions to the things you would expect in python, and typed model objects (Observation, Taxon, etc.) with full IDE autocompletion
• 🔒 Security: Keyring integration for secure credential storage
• 📗 Docs: Example requests, responses, scripts, and Jupyter notebooks to help get you started
• 🧪 Testing: A dry-run testing mode to preview your requests before potentially modifying data
• 💚 Responsible use: Follows the API Recommended Practices without extra configuration; caching and rate-limiting features reduce bandwidth usage, errors, and unexpected throttling

Supported Endpoints

Many of the most relevant API endpoints are supported, including:

• 📝 Annotations and observation fields
• 🆔 Identifications
• 💬 Messages
• 👀 Observations (multiple formats)
• 📷 Observation photos + sounds
• 📊 Observation histograms, observers, identifiers, life lists, and species counts
• 📍 Places
• 👥 Projects
• 🐦 Species
• 👤 Users

Quickstart

Here are usage examples for some of the most commonly used features.

First, install with pip:

pip install pyinaturalist

Then, import and create a client object. This will be our main interface to the API:

from pyinaturalist import *
client = iNatClient()

Note: If you are looking for the lower-level API functions (without the client class), see this page

Search observations

Let's start by searching for all your own observations. There are numerous fields you can search on, but we'll just use user_id for now:

>>> results = client.observations.search(user_id='my_username')

The full response consists of Observation objects with numerous attributes, but we can use pyinaturalist.pprint() to print out a condensed summary:

>>> pprint(results)
ID         Taxon                               Observed on   User     Location
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
117585709  Genus: Hyoscyamus (henbanes)        May 18, 2022  niconoe  Calvi, France
117464920  Genus: Omophlus                     May 17, 2022  niconoe  Galéria, France
117464393  Genus: Briza (Rattlesnake Grasses)  May 17, 2022  niconoe  Galéria, France
...

You can also get observation counts by species. On iNaturalist.org, this information can be found on the 'Species' tab of search results. For example, to get species counts of all your own research-grade observations:

>>> counts = client.observations.species_counts(user_id='my_username', quality_grade='research')
>>> pprint(counts)
 ID     Rank      Scientific name               Common name             Count
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
47934   species   🐛 Libellula luctuosa         Widow Skimmer           7
48627   species   🌻 Echinacea purpurea         Purple Coneflower       6
504060  species   🍄 Pleurotus citrinopileatus  Golden Oyster Mushroom  6
...

The data will be in the form of TaxonCount objects:

>>> counts[0]
TaxonCount(
    id=48662,
    name='Danaus plexippus',
    preferred_common_name='Monarch',
    rank='species',
    count=13,
    observations_count=458712,
    ...
)

Another useful format is the observation histogram, which shows the number of observations over a given time interval. The default is month_of_year:

>>> histogram = client.observations.histogram(user_id='my_username')
>>> pprint(histogram)
Month   Count
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Jan     8       ████
Feb     1       █
Mar     20      ██████████

The raw data will be a dict, with either int or datetime keys, depending on the interval:

>>> print(histogram.raw)
{
    1: 8,  # January
    2: 1,  # February
    3: 19, # March
    ...,   # etc.
}

Create and update observations: authentication

To create or modify observations, you will first need to log in. This requires creating an iNaturalist app, which will be used to get an access token.

creds = {
    'username': 'my_username',
    'password': 'my_password',
    'app_id': 'my_app_id',
    'app_secret': 'my_app_secret',
}
client = iNatClient(creds=creds)

See Authentication for more options including environment variables, keyrings, and password managers. A keyring is recommended, which does not require passing credentials directly:

client = iNatClient()
# Creds will be requested from the keyring when an authenticated request is made
client.observations.create(...)

Create and update observations

Now we can create a new observation:

from datetime import datetime

new_obs = client.observations.create(
    taxon_id=54327,  # Vespa Crabro
    observed_on_string=datetime.now(),
    time_zone='Brussels',
    description='This is a free text comment for the observation',
    tag_list='wasp, Belgium',
    latitude=50.647143,
    longitude=4.360216,
    positional_accuracy=50,  # GPS accuracy in meters
    photos=['~/observations/wasp1.jpg', '~/observations/wasp2.jpg'],
    sounds=['~/observations/recording.wav'],
)

We can then update the observation information, photos, or sounds:

client.observations.update(
    new_obs.id,  # Use the observation ID from the result above
    access_token=token,
    description='updated description !',
    photos='~/observations/wasp_nest.jpg',
    sounds='~/observations/wasp_nest.mp3',
)

Search species

There are many more resource types available besides observations. Taxonomy is another useful one.

Let's say you partially remember either a genus or family name that started with 'vespi'-something. The taxon search can be used to search by name, rank, and several other criteria:

>>> results = client.taxa.search(q='vespi', rank=['genus', 'family'])

As with observations, there is a lot of information available in the response (Taxon objects), but we'll print just a few basic details:

>>> pprint(results)
ID        Rank     Scientific name    Common name
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
52747     family   🐝 Vespidae        Hornets, Paper Wasps, Potter Wasps, and Allies
84737     genus    🦋 Vespina
646195    genus    🪰 Vespiodes
...

Next Steps

For more information, see:

• User Guide: introduction and general features that apply to most endpoints
• Endpoint Summary: a complete list of endpoints wrapped by pyinaturalist
• Examples: data visualizations and other examples of things to do with iNaturalist data
• Reference: Detailed API documentation
• Contributing Guide: development details for anyone interested in contributing to pyinaturalist
• History: details on past and current releases
• Issues: planned & proposed features

Feedback

If you have any problems, suggestions, or questions about pyinaturalist, you are welcome to create an issue or discussion. Also, PRs are welcome!

Note: pyinaturalist is developed by members of the iNaturalist community, and is not endorsed by iNaturalist.org or the California Academy of Sciences. If you have non-python-specific questions about the iNaturalist API or iNaturalist in general, the iNaturalist Community Forum is the best place to start.

Related Projects

Other python projects related to iNaturalist:

• naturtag: A desktop application for tagging image files with iNaturalist taxonomy & observation metadata
• pyinaturalist-convert: Tools to convert observation data to and from a variety of useful formats
• pyinaturalist-notebook: Jupyter notebook Docker image for pyinaturalist
• dronefly: A Discord bot with iNaturalist integration, used by the iNaturalist Discord server.