INDUSTRY FIELD GUIDE
Hydropower reporting intervals for AI evaluation
A water observation and a generation figure can have different reporting windows.
NAICS 221111 · Hydroelectric Power Generation
A water observation and a generation figure can have different reporting windows. EIA explains hydropower and publishes detailed electricity data. A benchmark can test whether an assistant recognizes which intervals can be compared before describing a relationship. EIA hydropower explanation · EIA Form 923 data
Create a fictional month with a plant-level total and several invented daily observations. Align one observation to the reporting interval and give another a different time zone or date range. Ask the assistant which figures may be compared, what the record actually shows, and why one observation cannot establish the cause of a change.
Use approved public aggregates under applicable terms or artificial values. Keep dam controls, exact infrastructure locations, reservoir operating information, security details and raw telemetry outside the sample. The output is retrospective data review.
Describe a hydro reporting example
Original Fiund exercise
Check observation membership in a stated reporting window
An original reporting exercise uses artificial values and a UTC month boundary. It tests time alignment, not hydropower operations or causal performance analysis.
| Record | Source | What it records |
|---|---|---|
| REPORT-H1 | Practice monthly definition | January covers January 1 at 00:00 UTC up to, but not including, February 1 at 00:00 UTC. |
| OBS-H2 | First practice observation | The timestamp is January 17 at 12:00 UTC. |
| OBS-H3 | Offset-clock observation | The timestamp is January 31 at 23:30 with UTC offset minus two hours. |
| TOTAL-H4 | Monthly practice aggregate | A generation total is labeled for H1's January window, without an explanation of any change. |
Question for the model
Which observation lies within H1's January window? Does membership explain the monthly generation figure?
Show the reference answer and checks
H2 lies within the stated January window. H3 converts to February 1 at 01:30 UTC, outside that window. A shared reporting interval would only establish alignment; it would not prove why a monthly figure changed or make a single observation representative of the whole month. The aggregate supplies no causal explanation.
- Assigning H3 to January from its displayed local date alone.
- Treating one aligned observation as a demonstrated cause of a monthly change.
Fields to preserve
- Window convention
- Preserve the included start and excluded end.
- Timestamp and offset
- Convert observations on their actual stated basis.
- Aggregation scope
- Keep a point observation distinct from a monthly total.
Test another case
Add an observation exactly at February 1 at 00:00 UTC. Under H1's rule it belongs outside January.
Use invented records or cleared public aggregates. Exclude dam controls, reservoir operating information, security details, real locations and raw telemetry.
Discuss a contribution
Describe the sources you control, the task they could support and what has already been checked. Use the existing-data assessment for an archive or the collection brief for a new recording project. The contributor agreement defines Fiund’s licensing authority and compensation. A guide or fictional exercise does not establish an accepted collection or a sale.