Time series are collections of values that are ordered in time. Preserving this ordering helps identify trends, detect seasonal patterns, and predict future values. Such series show up in many fields, from econometrics (unemployment rates, …), finance (stock prices, …), and demography (birth rates, …) to meteorology (rainfall, …), physiology (heart rates, …), and information technology (network traffic, …). Time series are tightly integrated into the Wolfram Language, allowing for seamless workflows with absolute or calendar time, regular or irregular sampling, scalar or vector values, single or multiple series, and in the presence of missing data. The Wolfram Language offers an extensive collection of tools for processing time series. These tools range from descriptive statistics, filters, and visualization to forecasts, simulation, and highly automated modeling frameworks.

Construction

TimeSeries — series of time-value pairs

EventSeries — special time series with no interpolation between samples

TemporalData — a collection of time series

ResamplingMethod  ▪  MissingDataMethod  ▪  TemporalRegularity

RandomFunction  ▪  FinancialData  ▪  FinancialIndicator

Country City Company Movie ...

Visualization »

DateListPlot — plot time series data

StackedDateListPlot — plot multiple time series data stacked on top of each other

DateListLogPlot  ▪  DateListStepPlot  ▪  TimelinePlot  ▪  DateHistogram  ▪  Histogram  ▪  ...

Basic Operations

TimeSeriesWindow — give the time series in the specified time window

TimeSeriesInsert — insert time-value pairs into a time series

TimeSeriesRescale  ▪  TimeSeriesResample  ▪  TimeSeriesShift  ▪  TimeSeriesThread  ▪  TimeSeriesMap  ▪  TimeSeriesMapThread  ▪  RegularlySampledQ  ▪  MinimumTimeIncrement

Basic Statistics »

Mean — find the mean of the values

StandardDeviation  ▪  Variance  ▪  Median  ▪  Quantile  ▪  ...

EmpiricalDistribution — find the empirical distribution of the values

HistogramDistribution  ▪  KernelMixtureDistribution  ▪  EstimatedDistribution

Filtering & Aggregating Time Series

MovingMap — apply a function to a moving overlapping window

TimeSeriesAggregate — apply a function to a moving non-overlapping window

Differences  ▪  Accumulate  ▪  MovingAverage  ▪  MovingMedian  ▪  ...

LowpassFilter  ▪  HighpassFilter  ▪  MeanFilter  ▪  ...

Fitting and Interpolation

FindFit — fit a function of time to the time series

Interpolation  ▪  LinearModelFit  ▪  NonlinearModelFit  ▪  ...

Time Series Process Modeling »

TimeSeriesModelFit — automatically fit a time series model

TimeSeriesForecast  ▪  CorrelationFunction  ▪  PowerSpectralDensity  ▪  ...

Time Specifications »

DateObject, TimeObject — date, time specifications

Now  ▪  DatePlus  ▪  LocalTime  ▪  TimeZoneConvert  ▪  UnitConvert  ▪  ...

Wolfram Data Drop »

Databin — representation of a databin accumulated in the Wolfram Data Drop