datescale

Datescale is a term used in data science and digital humanities to denote a continuous numeric representation of calendar dates, enabling arithmetic, comparisons, and statistical analysis of temporal data. The concept treats a date as a real number on a linear scale, with units such as days or seconds defined relative to a fixed epoch.

Common implementations use established epoch-based representations, including Unix time (seconds since 1970-01-01 UTC) and Julian Day Number (days since a fixed point in the proleptic Julian calendar). Some workflows also employ modified or custom epochs depending on dataset requirements.

Converting ISO 8601 date strings to a datescale requires handling time zones, daylight saving time, leap years, and leap seconds. Some scales represent only dates (no time of day), while others preserve sub-day precision.

Advantages include straightforward arithmetic on dates, easy sorting and alignment in time series, and compatibility with models that expect numerical inputs. Limitations involve calendar irregularities, potential loss of calendar semantics if not carefully represented, and precision issues when integrating diverse scales.

Usage appears in databases, data pipelines, and research that combine temporal data from multiple calendars. Because the term datescale is not standardized, it often describes a concept or a particular implementation rather than a formal specification.

See also: Unix time, Julian day, ISO 8601, Epoch, Time series, Chronology.