Docs
Scoring API: FAQ
title: "Scoring API FAQ"
Scoring API FAQ
What counts as a row?
Every non-header row in your CSV counts as one row, whether it scores successfully or contains invalid values. The credit charge is based on the total number of data rows in the file: 1 credit per 100 rows, rounded up. A file with 1 row costs 1 credit. A file with 101 rows costs 2 credits.
How long are download links valid?
Download URLs for scored results expire after 24 hours. If you need the results after that, upload the file again. The same input will produce the same output.
Can I score the same data twice?
Yes. Scoring is deterministic. The same input rows will always receive the same profile assignments, so re-scoring is safe. You will be charged credits for each request.
Do I need to register an app or create an OAuth client to use the API?
No. The API uses a simple API key in the Authorization header. There is no OAuth flow, no client registration, and no callback URLs. Generate a key from your dashboard and start making requests.
What happens to my data after scoring?
Your uploaded CSV is processed and deleted within 24 hours. The scored output CSV is available via a presigned download URL that also expires after 24 hours. AUSynth does not store your data permanently, does not use it to retrain the profiling model, and does not share it with anyone. Logs retain job metadata (timestamps, row counts, credits) but not your input data.
Can I use raw ABS Census labels instead of the collapsed ones?
Yes. If your data already uses ABS labels (for example, 25-29 years instead of Early career), the API maps them automatically. You can also mix both formats within the same file. See Variable Reference for the full mapping.
What if some of my data is missing or has values the API does not recognise?
The API scores every row regardless. Unrecognised values are mapped to Not stated and the row is still assigned a profile. If three or more of the nine variables are Not stated, the row receives profile_confidence: "low" instead of "high". The response includes a data_quality object listing which rows had unrecognised values so you can fix your data for future uploads.
How fast is scoring?
Files under 10,000 rows typically return in under a second. Files between 10,000 and 100,000 rows complete within 10 seconds. A million rows takes about a minute. Files over 10,000 rows are processed asynchronously; the API returns a poll_url that you check until scoring is complete.
What are the eight profiles?
The profiles are derived from nine demographic and socioeconomic variables using statistical clustering. They are:
| ID | Name |
|---|---|
| 0 | Labourers and operators |
| 1 | Young singles and non-workers |
| 2 | Children |
| 3 | Non-earning dependants |
| 4 | Trades and technical workers |
| 5 | Established partnered households |
| 6 | Retired and semi-retired |
| 7 | High-earning professionals |
Profile assignments are deterministic and stable. The same input always produces the same profile.
How many credits do I get on the free tier?
Free accounts receive 5 credits per week. Credits are unified across all AUSynth products including data downloads, analyses, and scoring. Five credits is enough to score up to 500 rows per week.
Does scoring use a different credit pool from data downloads?
No. All AUSynth credits come from the same pool. If you buy a 500-credit bundle, you can spend them on data downloads, analysis reports, or scoring in any combination.