Integrating with Pandas

The OpenAQ Python SDK is designed around native Python data structures and in turn does not return data as vectorized arrays such as data frames used by packages like Numpy, Pandas and others. Because native Python data structures are used integrating these types of libraries is straightforward.

Because the data returned from the OpenAQ API is highly nested as a deserialized Python object we need to flatten the resulting response so that it fits as a 2-dimensional array in Pandas. Fortunarely, Pandas provides a function to facilitate this, json_normalize:

SyncAsync

from openaq import OpenAQ
from pandas import json_normalize

client = OpenAQ(api_key='replace-with-a-valid-openaq-api-key')
response = client.locations.list(
    bbox=[-0.464172,5.449908,0.030212,5.691491]
)
data = response.dict()
df = json_normalize(data['results'])
client.close()

import asyncio

from openaq import AsyncOpenAQ
from pandas import json_normalize

async def main():
    client = AsyncOpenAQ(api_key='replace-with-a-valid-openaq-api-key')
    response = await client.locations.list(
        bbox=[-0.464172,5.449908,0.030212,5.691491]
    )
    data = response.dict()
    df = json_normalize(data['results'])
    await client.close()

if __name__ ==  '__main__':
    loop = asyncio.get_event_loop()
    loop.run_until_complete(main())