ausdata — Python SDK for the Australian public data API

The official Python client for ausdata.io. One auth key, one uniform response envelope, every major Australian government data source.

pip install ausdata-sdk
# or, with pandas convenience:
pip install "ausdata-sdk[pandas]"

Why?

The chart you spend 90 minutes building in Excel — one HTTP call.

# Before — wrangling SDMX XML by hand
curl 'https://api.data.abs.gov.au/data/WPI/...'

# After — typed, normalised, attribution-tagged

from ausdata import Client

client = Client(api_key="ak_xxx")
result = client.real_wages(start="2019-Q1", end="2024-Q4")

print(result.data[0])
# RealWageRow(period='2024-Q4', wpi_annual_change_pct=3.2,
#             cpi_annual_change_pct=2.4, real_wages_gap_pct=0.8,
#             real_wages_direction='growing')

df = result.to_dataframe()       # pandas (optional extra)
result.to_csv("real_wages.csv")  # Datawrapper-friendly

Async client

import asyncio
from ausdata import AsyncClient

async def main():
    async with AsyncClient(api_key="ak_xxx") as client:
        snapshot = await client.economic_dashboard()
        print(snapshot.data.cash_rate_pct, snapshot.data.unemployment_rate_pct)

asyncio.run(main())

Endpoints

The full discover → introspect → fetch loop is:

r = client.search_datasets(q="unemployment rate")  # find dataset
schema = client.describe("abs", "LF")              # see valid filters
data = client.get_data("abs", "LF",                # fetch with right filters
                       measure="unemployment_rate",
                       region="nsw", start="2024-01")

Every data response is wrapped in the same ApiResponse envelope so you always know where to find the payload (.data), the attribution (.meta.sources), and the audit trail (.meta.retrieved_at, .meta.stale).

Pricing

Sign up + get a free key at ausdata.io.

Configuration

from ausdata import Client

# API key resolution order:
#   1. explicit api_key= argument
#   2. AUSDATA_API_KEY environment variable
client = Client(
    api_key="ak_xxx",
    base_url="https://api.ausdata.io",   # override for staging
    timeout=30.0,                         # per-request timeout (seconds)
    max_retries=3,                        # 5xx + 429 retry budget
    retry_backoff_factor=2.0,             # exponential delay multiplier
)

Errors

from ausdata import Client
from ausdata.exceptions import (
    AusdataError,         # base
    AuthenticationError,  # 401
    PermissionError,      # 403, 402 (tier-blocked)
    RateLimitError,       # 429 with retry_after
    ValidationError,      # 400
    UpstreamError,        # 502/503
    AusdataServerError,   # 5xx fallback
)

try:
    result = client.real_wages(start="not-a-quarter")
except ValidationError as e:
    print(e.message, "Hint:", e.hint)

Every error carries the API's message plus the actionable hint when available (e.g. "Did you mean 2024-Q1?").

Links

• Docs: docs.ausdata.io
• API reference: api.ausdata.io/docs
• Issues: github.com/Bigred97/ausdata-sdk/issues

License

MIT. See LICENSE.