Create a time series from time-value pairs:

Create a time series from values and times:

Specify keys for the value components:

Specify value keys with an option:

Use iterator to specify timestamps:

With multivariate values:

Create a time series with Automatic timestamps:

The default timestamps are non-negative integers:

Extract values:

Select single value component:

Single value component tabular:

Select multiple value components:

Repeated components will be discarded since keys must be unique in a Tabular object:

Create a time series with values given as a Tabular object:

Extract timestamps:

Get the type information for the timestamps:

Extract a sub time series for a given temporal window:

Plot a time series as a function of time:

Give explicit component keys:

Show the data using conversion to Tabular:

Component keys are added automatically when the given list is incomplete:

Generate all component keys by giving {}:

The Automatic value of the component keys argument preserves the input keys:

When component keys are given, they override the input keys:

With component keys value None, the output time series is simple, without explicit components:

When there are no input keys, the Automatic value will preserve the simple component input:

Numerical, listable functions automatically thread over values of time series:

Compare to the result of TimeSeriesMap:

Find the Mean of a time series:

The mean depends only on the values:

Compute a weighted mean weighted by distribution of timestamps:

Combining several time series with identical timestamps threads over values:

Time series are resampled at the union of timestamps in the intersection of their supports:

The intersection of supports is , and the union of the timestamps within is :

Compare the result with adding the resampled time series:

Create a new time series of quantity magnitudes from existing time series involving quantities:

Compute the difference between two time series:

Compare the result with computation using TimeSeriesThread:

Visualize a time series:

Use TimeSeriesWindow to extract a portion of a time series:

Use TimeSeriesInsert to replace a missing value:

Use TimeSeriesShift to shift the series ahead by 2:

Use TimeSeriesRescale to stretch a time series:

Use TimeSeriesMap to find the sums of the components of a vector-valued time series:

Compute a MovingAverage for a time series:

Use MovingMap to compute a moving maximum of a time series:

Use TimeSeriesAggregate to compute the weekly totals for a time series:

Take a time series sampled over powers of 2:

Upsample the original path in steps of 2:

The new data is sampled from the path function:

Fit a parametric model to a time series using TimeSeriesModelFit:

Use TimeSeriesForecast to forecast the next 10 values in the time series:

Plot the time series along with the forecast:

By default, dates without an explicit calendar specification are interpreted as "Gregorian":

Using CalendarTypecal will change those "Gregorian" dates to the calendar cal:

Construct a time series with dates in the "Julian" calendar:

Convert those dates to the "Islamic" calendar:

When dates are given in different calendars, they will be converted into the same calendar:

Use the CalendarType option to convert to a specific calendar:

By default, data without explicit components results in a simple time series:

Show the data using conversion to Tabular:

Construct a time series with components keys "a" and "b":

By default, the date granularity is inferred from the input timestamps:

Use DateFunction to select a different granularity:

Take a list of date-value pairs with ambiguous dates:

Specify a year-month-day interpretation:

Specify the alternative year-day-month interpretation:

Specify the date granularity of a TimeSeries:

Requiring a finer date granularity will instantiate the beginning of the timestamps:

Specify the holiday calendar to use with business days:

The fourth of July is not a holiday in Mexico:

Federal business days may be different from those of a stock exchange:

Use HolidayCalendar to visualize business days in a given country:

By default, TimeSeries interpolates between timestamps using linear order:

Specify quadratic interpolation:

By default, TimeSeries takes the mean value for duplicate timestamps:

Take the total of the corresponding values:

Include additional metainformation:

Extract the metainformation:

The properties now include the meta-information "Stock":

The added meta-information can be used like any other property:

Take a time series with missing values:

Use MissingDataMethod to replace missing values with a mean of non-missing values:

Use linear interpolation to replace the missing values:

Specify interpolation order:

By default, values at intermediate times in a TimeSeries are computed using linear interpolation:

Use second-order interpolation:

Explicitly assume that temporal data is regularly spaced:

By default, the time observation window is inferred from the range of the timestamps:

Specify a different time observation window:

Specify the time zone of a TimeSeries:

Date string timestamps are assumed to be given in $TimeZone, even with TimeZone specified:

The TimeZone option affects the timestamps but not the values, even if they are dates:

Use SunPosition to generate the Sun's position in Chicago for a range of dates:

Plot the variation of the azimuth and the altitude for this period:

Use MoonPosition to generate the Moon's position for a range of dates:

Verify that the Moon's orbit is tilted with respect to the Earth's equator:

The number of hours of sunlight per day in 2015 for Champaign, IL:

Find the mean length of sunlight per day:

The change of sunlight length daily in minutes:

Use CountryData to generate the GDP for UK and Germany:

Compare the GDP of these two countries:

Forecast stock prices:

Fit an ARIMA process:

Forecast to the next half a year:

Use AirPressureData to examine pressure reading drops due to Hurricane Sandy at Long Island MacArthur Airport:

Use WindSpeedData to compare the wind speeds at John F. Kennedy Airport during summer and winter:

Average temperature on the first day of a month in Chicago, IL:

Resample to get uniformly spaced time series:

Fit a SARIMA process:

Use NuclearReactorData to visualize energy production for the Chernobyl reactors:

Capture a 10-second time series at 0.05-second intervals:

Plot the time series along with a moving average:

The following time series is generated by reading illuminance data from a TinkerForge Weather Station every 0.1 second for 5 seconds while continuously changing the device orientation:

Plot the illuminance values:

Min, Mean and Max for the data:

Generate illuminance data by alternately switching the ambient light source on and off:

A plot of the time series reveals the approximate periodic nature of the data:

Verify the periodicity using Fourier:

Use MeanFilter to filter a time series:

The following data represents annual sales for a small software company for 11 years:

Use LinearModelFit to fit a linear model to this data:

Plot the original data along with the values obtained using the linear model:

Apply exponential smoothing with weight 0.45 to the data:

Plot the original data along with the values obtained using exponential smoothing:

Fit a parametric model to retail sales data for the US between 1992 and 2015:

Construct a time series model for the data using TimeSeriesModelFit:

Use the model to forecast the next three months: