Seasonal aware method for forecasting and capacity planning

One or more non-transitory computer-readable media storing instructions, wherein the instructions include:

• instructions, which when executed by one or more hardware processors, cause associating a particular class of time-series data from a set of time-series data with at least two sub-classes of data;
• wherein the set of time-series data comprises a measurement of at least one value, and wherein the at least two sub-classes of data are determined based at least in part on measured differences in the at least one value;
• wherein a first sub-class of data from the at least two sub-classes of data classifies first-subclass member data as having high values relative to at least part of the set of time-series data, and a second sub-class of data from the at least two sub-classes of data classifies second-subclass member data at least as having non-high values relative to at least part of the set of time-series data;
• instructions, which when executed by one or more hardware processors, cause determining a trend rate for a particular sub-class from the first sub-class or the second sub-class of data;
• instructions, which when executed by one or more hardware processors, cause generating, based at least in part on the trend rate for the particular sub-class of data, a forecast that projects future values for the set of time-series data that belong to the particular sub-class of data.