You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Although ERCOT clearly delineates the supply of both medium-term and long-term forecast products, NYISO is less explicit in their offerings. It seems that they similarly offer a medium-term forecast product (6-day) and a long-term forecast product, for capacity planning (ERCOT does not have a capacity market).
On the load data page, there are three sections:
NY Load Forecast
Zonal Load Commitment
Load Forecast Weather Data
Our focus is on 1 here. Unfortunately the files do not contain publish times, and it seems like each time a new forecast is issued it will be overwritten, unless they are written only once daily. As of the time of writing (1:30pm EDT on August 17th), the system operator claims the most recently issued forecast is 8/17 at 7:45pm EDT for August 18th, which is impossible because that is in the future, unless they claim that will be the timestamp of the final issuance of the forecast. In that file, hourly forecasts are issued for the next 6 days (August 18th-23rd) for 11 zones plus the system.
EDIT: I'm dumb and forgot there are 24 hours in a day, and 07:45 likely refers to the morning.
Thus, the previous 5 files (8/17, 8/16, ...) contain the previously issued forecasts. Without being explicitly stated, we must assume that these load forecasts are issued once daily.
Unfortunately the behavior of load forecast for NYISO is also currently broken in the gridstatusio client. The figure below shows that the publish time (red) is after most of the operating day, which doesn't corroborate with the behavior described on the website (the final forecast is issued the day before the operating day).
There is a historical archive of CSV files found at http://mis.nyiso.com/public/P-7list.htm, which can be used to backfill (currently the NYISO load forecast scraper only goes back to 2023-04-12.
Zonal Load Commitment details for each hour the amount of bids and offers in each zone. The publishing time behavior is similar to 1.
Load Forecast Weather Data is temperature bands with Actual and Forecast vintage types for each zone. This means it is updated ex post facto.
The open questions are:
how do we know how many forecasts are issued daily? (sent email, awaiting reply)
how do we fix the behavior of the gridstatusio client to return a coherent publish time?
do we want to be able to allow people to grab forecasts from more than one day out? How would we allow for this functionality and parametrization?
The scraper's behavior seems to be working properly, although the Forecast Time is simply set as the date of the file. It's unclear, still, when the forecasts are actually published (i.e. how much lead time).
The text was updated successfully, but these errors were encountered:
NYISO supplies load forecast data here: https://www.nyiso.com/load-data.
Although ERCOT clearly delineates the supply of both medium-term and long-term forecast products, NYISO is less explicit in their offerings. It seems that they similarly offer a medium-term forecast product (6-day) and a long-term forecast product, for capacity planning (ERCOT does not have a capacity market).
On the load data page, there are three sections:
Our focus is on 1 here. Unfortunately the files do not contain publish times, and it seems like each time a new forecast is issued it will be overwritten, unless they are written only once daily. As of the time of writing (1:30pm EDT on August 17th), the system operator claims the most recently issued forecast is 8/17 at 7:45pm EDT for August 18th, which is impossible because that is in the future, unless they claim that will be the timestamp of the final issuance of the forecast. In that file, hourly forecasts are issued for the next 6 days (August 18th-23rd) for 11 zones plus the system.
EDIT: I'm dumb and forgot there are 24 hours in a day, and 07:45 likely refers to the morning.
Thus, the previous 5 files (8/17, 8/16, ...) contain the previously issued forecasts. Without being explicitly stated, we must assume that these load forecasts are issued once daily.
Unfortunately the behavior of load forecast for NYISO is also currently broken in the
gridstatusio
client. The figure below shows that the publish time (red) is after most of the operating day, which doesn't corroborate with the behavior described on the website (the final forecast is issued the day before the operating day).There is a historical archive of CSV files found at http://mis.nyiso.com/public/P-7list.htm, which can be used to backfill (currently the NYISO load forecast scraper only goes back to 2023-04-12.
The open questions are:
The scraper's behavior seems to be working properly, although the Forecast Time is simply set as the date of the file. It's unclear, still, when the forecasts are actually published (i.e. how much lead time).
The text was updated successfully, but these errors were encountered: