Platform, API and Client Library Feedback

[nc9]

We're steadily allowing user to access the OpenElectricity Platform.

If you're not yet on the waitlist - you can signup at:

https://platform.openelectricity.org.au

The client libraries are at:

Typescript - https://github.com/opennem/openelectricity-typescript

Python - https://github.com/opennem/openelectricity-python

You can file issues in each respective client library, or offer suggestions here. There are still more features to build out in the API and the client libraries, but happy to hear feedback and suggestions!

[ArthurBond]

Hi, thanks for making this available. I was just wondering whether renewable_proportion is available soon for NEM and WEM network regions? I am getting an 'Metric Metric.RENEWABLE_PROPORTION not supported' error.

[nc9]

We disabled it because we're formalising a definition of demand. You can track the progress of the open issue at #398

Just one note - expect that the endpoint may be in a bit of state of flux - I'll keep the clients updated as often as I can, but we're hoping to lock the interface down in a backward-compatible sense by the end of this week.

[ArthurBond]

Ah I see, thanks for the quick response. Do you have a rough idea of when it will be formalised?

I'm not sure what the emissions intensity calculation you are trying to do with the third definition of demand, but a quick inspection suggests that OPERATIONAL_DEMAND is greater or equal to TOTALDEMAND + DISPATCHABLELOAD (roughly speaking). In the MMS Data Model Report it also mentions that DISPATCHABLELOAD can be added to TOTALDEMAND to get a sense of the inherent demand in each region. It could be interesting to not subtract battery charging and pumps to capture the effect of their emissions as well (assuming they are dispatchable loads which batteries certainly are).

Then again, it probably makes more sense to subtract it if you are looking at the emissions coming from the generation of electricity.

I'm sure OE has already thought of this but I hope a second opinion helps!

[dianyo]

Hi @nc9,
I had some more general feedback for the API usage. I sometimes get internal server error when using API call, even using the same API which is workable before. I guess that might come from dev / stg environment not properly divided or regression test not been done properly when deploy a new version. Better to have those things for the platform to make things more robust. I can do some help if you want!

[npellow123]

I am trying to get the price data via the API:

curl -G "https://api.openelectricity.org.au/v4/data/network/NEM" \ -H "Authorization: Bearer ${API_KEY}" \ -d metrics='price' \ -d interval=5m \ -d date_start=2025-06-24T00:00:00 \ -d date_end=2025-07-01T23:59:59

However get the following response

{"version":"4.0.14","response_status":"ERROR","error":"Metric Metric.PRICE not supported","success":false}

Any idea when PRICE might be supported ?

[trainman75]

I found the following string worked for me (as used within a Home Assistant automation, so the format is a little different):
curl --request GET \ --url 'https://api.openelectricity.org.au/v4/market/network/NEM?metrics=price&interval=5m&with_clerk=true&primaryGrouping=network_region&network_region=VIC1&date_start=2025-07-15T00:00:00&date_end=2025-07-17T09:00:00' \ --header 'Authorization: Bearer {API_KEY}'

Key items to note:

• Change v4/data to v4/market
• Drop the quotes around 'price'

Although of note this morning is that data seems to have stopped around midnight (separate issue)

[ChrisC28]

Hi,
I'm looking to extract historical data (as far back in time as possible) at an output interval of (ideally) hourly or daily. With the current API settings, only the last year of data is available.

Is there any likelihood of access to the whole historical archive of NEW and WEM data being supported in the near future?

[mdavis-xyz]

I'm confused by this section of the Python docs.

The SDK provides built-in support for converting responses to Pandas and Polars DataFrames.

import pandas as pd

# Get market data
response = client.get_market(
    network_code="NEM",
    metrics=[MarketMetric.PRICE, MarketMetric.DEMAND],
    interval="1h",
    date_start=datetime(2024, 1, 1),
    date_end=datetime(2024, 1, 2),
    primary_grouping="network_region"
)

# Convert to DataFrame
data = []
for timeseries in response.data:
    for result in timeseries.results:
        for data_point in result.data:
            data.append({
                "timestamp": data_point.timestamp,
                "metric": timeseries.metric,
                "value": data_point.value,
                "unit": timeseries.unit
            })

df = pd.DataFrame(data)

# Analyze
print(df.groupby("metric")["value"].describe())

To me this is the opposite of "built-in support". If I have to write/copy-paste nested for-loops myself, then it is not built in.

I would consider a "built-in" approach to be:

response = client.get_market()
df = pl.DataFrame(response.for_polars())

or

response = client.get_market()
df = response.to_polars()

or

df = client.get_market(format="polars")

I would find such an approach more useful. In particular, I'm sure I will modify the list of metrics in the API call, and then forget to add another line in the dictionary.
Additionally, this example with for loops seems incomplete. I would have to additionally call .pivot() to get it into the kind of dataframe shape I expect, and that is a bit fiddly.

Lacking that kind of functionality, it think it would be more clear to say in the documentation that the library does not currently have built-in support for pandas/polars, but that you can cast to a DataFrame via native lists of dicts and then pivot.

[mdavis-xyz]

Ah, it seems the functionality I want does exist, according to the README. I think the other part of the documentation is stale.

https://github.com/mdavis-xyz/openelectricity-python?tab=readme-ov-file#using-with-pandas

df = response.to_pandas()

[luciaxu2010]

Hi all, I’m having trouble retrieving historical data older than around 2025-06 using get_network_data() via the Python client/API at the moment.

For example, even a minimal call like:

client.get_network_data(
    network_code="NEM",
    metrics=[DataMetric.POWER],
    interval="5m",
    date_start=datetime(2019, 1, 1),
    date_end=datetime(2019, 1, 2),
)

returns 400 Bad Request, while recent dates work fine.

The documentation mentions:

NEM data since: 1998-12-06
WEM data since: 2006-09-19

so I’m wondering whether:

• there is currently any historical availability limitation for general/public users,
• some metrics/groupings are only available for recent periods,
• or whether this may be related to the recent backend/NEMWeb migration.

Any clarification would be much appreciated, thanks!

[nc9]

you've likely run into request limits here - at per-interval the most you can get out is 7 days as the returns are large. check this page:

https://docs.openelectricity.org.au/api-reference/data-limits

let me know if there are any issues once you adjust

[luciaxu2010]

Hey Nik, thanks for your reply. Yes, I’m aware that hourly data needs to be fetched in monthly chunks via a loop. However, from my testing, it seems the data availability stops around 2025-06-ish, even when querying earlier historical ranges. Just wondering if there is currently any backend limitation or incomplete archive issue for public users?

[nc9]

ahh yes! sorry just saw the date. I had to trim back a lot of limits since we were getting hot spots from queries on our servers. can you email me on nik at opennem.org.au and let me know what data you need? we'll sort something out