Interval Data

Overview and Statuses

Keystone Designer requires a one year of continuous 15-minute interval data in order to process a scenario. Hourly data can also be used if 15-minute interval data is not available. Data must be uploaded from your local file system in a single file and can be specified in one of two formats:

• Green Button XML
• 2-column CSV, where the columns are:
  • Column 1: Interval timestamp in the project’s local timezone
  • Column 2: instantaneous load in kW.

A valid file template for CSV data is available for download by clicking Download 2 Column CSV Template

Energy use profiles can be in one of the following statuses:

• Processing: a new profile has been saved, and the interval data is being uploaded, parsed, and validated.
• Invalid: the interval data provided for the profile does not meet our requirements; specific details about the validation issues are provided by clicking on the invalid status indicator.
• Valid: the profile has been processed and is ready for use within an optimization scenario

NOTE: our internal processing of interval data may shift the provided date range of the raw interval data, so dates shown in the UI may not match exactly the dates provided in the source file.

Profiles can be removed from a project by clicking Delete under the action menu () if they are not associated with an optimization scenario.

Validation and Processing

When validating an interval data file, Keystone Designer is forgiving of occasional missing or invalid data points. Specifically, Keystone Designer will automatically handle:

Data gaps up to a week long

In this case Keystone Designer will use an average value for the specific missing timestamp/day of week combination from the equivalent timestamp/day-of-week values available in the rest of the file. For example, if Monday, October 10 from 10:00 - 12:00 is missing, Keystone Designer will fill that gap interval by interval, first taking an average of all available values for other Mondays at 10:00, then 10:15, and so on. We will do this as long as there is at least one value for that timestamp/day-of-week combination available in the file.

Invalid time stamps

Invalid timestamps and values at those timestamps will be ignored, and that resultant gap will be filled via the average method detailed above. All timestamps in a single file must be in either US or EU date format; date formats cannot be mixed. Valid timestamp formats include:

• mm/dd/yyyy hh:mm [AM/PM]
• dd/mm/yyyy hh:mm [AM/PM]

If AM/PM is not specified, we assume 24-hour time formatting.

Invalid values

Negative or empty kW values will be ignored, and that resultant gap will be filled via the average method detailed above.

Once we have processed the file, discarding invalid information and filling gaps, we validate that we have at least 12 months of contiguous 15-minute data. If so, the file is valid, and we further process the information by:

• Taking the latest 12 months from the uploaded file
• Realigning the data such that the resultant time series starts at midnight local time of the first of the closest month. For example, if the file starts at 10/10/2019 00:00 and ends at 10/09/2020 23:45, we will adjust the last nine days of the file to fill in the beginning of the month at the start of the file to result in a timeseries from 10/01/2019 00:00 - 9/30/2020 23:45.