Stats

Namespace: Deedle

The Stats type contains functions for fast calculation of statistics over series and frames as well as over a moving and an expanding window in a series.

The resulting series has the same keys as the input series. When there are no values, or missing values, different functions behave in different ways. Statistics (e.g. mean) return missing value when any value is missing, while min/max functions return the minimal/maximal element (skipping over missing values).

Remarks

The windowing functions in the Stats type support calculations over a fixed-size windows specified by the size of the window. If you need more complex windowing behavior (such as window based on the distance between keys), different handling of boundary, or chunking (calculation over adjacent chunks), you can use chunking and windowing functions from the Series module such as Series.windowSizeInto or Series.chunkSizeInto.

Expanding windows
Frame statistics
Moving windows
Multi-level statistics
Series interoploation
Series statistics

Expanding windows

Expanding window means that the window starts as a single-element sized window and expands as it moves over the series. In this case, statistics is calculated for all values up to the current key. This means that the result is attached to the key at the end of the window. The function names are prefixed with expanding.

Static members

Static member	Description
`expandingCount(series)` Signature: series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains counts over expanding windows (the value for a given key is calculated from all elements with smaller keys).
`expandingKurt(series)` Signature: series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains kurtosis over expanding windows (the value for a given key is calculated from all elements with smaller keys); If the entire window contains fewer than 4 values, the result is missing.
`expandingMax(series)` Signature: series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains maximum over an expanding window. The value for a key k in the returned series is the maximum from all elements with smaller keys.
`expandingMean(series)` Signature: series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains means over expanding windows (the value for a given key is calculated from all elements with smaller keys); If the entire window contains no values, the result is missing.
`expandingMin(series)` Signature: series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains minimum over an expanding window. The value for a key k in the returned series is the minimum from all elements with smaller keys.
`expandingSkew(series)` Signature: series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains skewness over expanding windows (the value for a given key is calculated from all elements with smaller keys); If the entire window contains fewer than 3 values, the result is missing.
`expandingStdDev(series)` Signature: series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains standard deviation over expanding windows (the value for a given key is calculated from all elements with smaller keys); If the entire window contains fewer than 2 values, the result is missing.
`expandingSum(series)` Signature: series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains sums over expanding windows (the value for a given key is calculated from all elements with smaller keys); If the entire window contains no values, the result is 0.
`expandingVariance(series)` Signature: series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains variance over expanding windows (the value for a given key is calculated from all elements with smaller keys); If the entire window contains fewer than 2 values, the result is missing.

Frame statistics

The standard functions are exposed as static members and are overloaded. This means that they can be applied to both Series<'K, float> and to Frame<'R, 'C>. When applied to data frame, the functions apply the statistical calculation to all numerical columns of the frame.

Static members

Static member	Description
`count(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,int> Type parameters: 'R, 'C	For each column, returns the number of the values in the column. This excludes missing values and values created from `Double.NaN` etc.
`kurt(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,float> Type parameters: 'R, 'C	For each numerical column, returns the kurtosis of the values in a series. The function skips over missing values and `NaN` values. When there are less than 4 values, the result is NaN.
`max(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,float> Type parameters: 'R, 'C	For each numerical column, returns the maximal values as a series. The function skips over missing and `NaN` values. When there are no values, the result is `NaN`.
`mean(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,float> Type parameters: 'R, 'C	For each numerical column, returns the mean of the values in the column. The function skips over missing values and `NaN` values. When there are no available values, the result is NaN.
`median(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,float> Type parameters: 'R, 'C	For each numerical column, returns the median of the values in the column.
`min(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,float> Type parameters: 'R, 'C	For each numerical column, returns the minimal values as a series. The function skips over missing and `NaN` values. When there are no values, the result is `NaN`.
`skew(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,float> Type parameters: 'R, 'C	For each numerical column, returns the skewness of the values in a series. The function skips over missing values and `NaN` values. When there are less than 3 values, the result is NaN.
`stdDev(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,float> Type parameters: 'R, 'C	For each numerical column, returns the standard deviation of the values in the column. The function skips over missing values and `NaN` values. When there are less than 2 values, the result is NaN.
`sum(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,float> Type parameters: 'R, 'C	For each numerical column, returns the sum of the values in the column. The function skips over missing values and `NaN` values. When there are no
`variance(frame)` Signature: frame:Frame<'R,'C> -> Series<'C,float> Type parameters: 'R, 'C	For each numerical column, returns the variance of the values in the column. The function skips over missing values and `NaN` values. When there are less than 2 values, the result is NaN.

Moving windows

Moving window means that the window has a fixed size and moves over the series. In this case, the result of the statisitcs is always attached to the last key of the window. The function names are prefixed with moving.

Static members

Static member	Description
`movingCount(size series)` Signature: size:int -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains counts over a moving window of the specified size. The first `size-1` elements of the returned series are always missing; if the entire window contains missing values, the result is 0.
`movingKurt(size series)` Signature: size:int -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains kurtosis over a moving window of the specified size. The first `size-1` elements of the returned series are always missing; if the entire window contains missing values, the result is also missing.
`movingMax(size series)` Signature: size:int -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains maximum over a moving window of the specified size. The first `size-1` elements are calculated using smaller windows spanning over `1 .. size-1` values. If the entire window contains missing values, the result is missing.
`movingMean(size series)` Signature: size:int -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains means over a moving window of the specified size. The first `size-1` elements of the returned series are always missing; if the entire window contains missing values, the result is also missing.
`movingMin(size series)` Signature: size:int -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains minimum over a moving window of the specified size. The first `size-1` elements are calculated using smaller windows spanning over `1 .. size-1` values. If the entire window contains missing values, the result is missing.
`movingSkew(size series)` Signature: size:int -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains skewness over a moving window of the specified size. The first `size-1` elements of the returned series are always missing; if the entire window contains missing values, the result is also missing.
`movingStdDev(size series)` Signature: size:int -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains standard deviations over a moving window of the specified size. The first `size-1` elements of the returned series are always missing; if the entire window contains missing values, the result is also missing.
`movingSum(size series)` Signature: size:int -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains sums over a moving window of the specified size. The first `size-1` elements of the returned series are always missing; if the entire window contains missing values, the result is 0.
`movingVariance(size series)` Signature: size:int -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Returns a series that contains variance over a moving window of the specified size. The first `size-1` elements of the returned series are always missing; if the entire window contains missing values, the result is also missing.

Multi-level statistics

For a series with multi-level (hierarchical) index, the functions prefixed with level provide a way to apply statistical operation on a single level of the index. (For example you can sum values along the 'K1 keys in a series Series<'K1 * 'K2, float> and get Series<'K1, float> as the result.)

Static members

Static member	Description
`levelCount(level series)` Signature: (level:('K -> 'L)) -> series:Series<'K,'V> -> Series<'L,int> Type parameters: 'K, 'L, 'V	For each group with equal keys at the level specified by `level`, returns the number of the values in the group. This excludes missing values and values created from `Double.NaN` etc.
`levelKurt(level series)` Signature: (level:('K -> 'L)) -> series:Series<'K,float> -> Series<'L,float> Type parameters: 'K, 'L	For each group with equal keys at the level specified by `level`, returns the kurtosis of the values in a series. The function skips over missing values and `NaN` values. When there are less than 4 values, the result is NaN.
`levelMean(level series)` Signature: (level:('K -> 'L)) -> series:Series<'K,float> -> Series<'L,float> Type parameters: 'K, 'L	For each group with equal keys at the level specified by `level`, returns the mean of the values in the group. The function skips over missing values and `NaN` values. When there are no available values, the result is NaN.
`levelMedian(level series)` Signature: (level:('K -> 'L)) -> series:Series<'K,float> -> Series<'L,float> Type parameters: 'K, 'L	For each group with equal keys at the level specified by `level`, returns the median of the values in the group.
`levelSkew(level series)` Signature: (level:('K -> 'L)) -> series:Series<'K,float> -> Series<'L,float> Type parameters: 'K, 'L	For each group with equal keys at the level specified by `level`, returns the skewness of the values in a series. The function skips over missing values and `NaN` values. When there are less than 3 values, the result is NaN.
`levelStdDev(level series)` Signature: (level:('K -> 'L)) -> series:Series<'K,float> -> Series<'L,float> Type parameters: 'K, 'L	For each group with equal keys at the level specified by `level`, returns the standard deviation of the values in the group. The function skips over missing values and `NaN` values. When there are less than 2 values, the result is NaN.
`levelSum(level series)` Signature: (level:('K -> 'L)) -> series:Series<'K,float> -> Series<'L,float> Type parameters: 'K, 'L	For each group with equal keys at the level specified by `level`, returns the sum of the values in the group. The function skips over missing values and `NaN` values. When there are no available values, the result is 0.
`levelVariance(level series)` Signature: (level:('K -> 'L)) -> series:Series<'K,float> -> Series<'L,float> Type parameters: 'K, 'L	For each group with equal keys at the level specified by `level`, returns the variance of the values in the group. The function skips over missing values and `NaN` values. When there are less than 2 values, the result is NaN.

Series interoploation

Static members

Static member Description

Static member	Description
`interpolate(keys f series)` Signature: keys:seq<'K> -> (f:('K -> ('K * 'T) option -> ('K * 'T) option -> 'T)) -> series:Series<'K,'T> -> Series<'K,'T> Type parameters: 'K, 'T	Interpolates an ordered series given a new sequence of keys. The function iterates through each new key, and invokes a function on the current key, the nearest smaller and larger valid observations from the series argument. The function must return a new valid float. Parameters `keys` - Sequence of new keys that forms the index of interpolated results `f` - Function to do the interpolating
`interpolateLinear(keys keyDiff series)` Signature: keys:seq<'K> -> (keyDiff:('K -> 'K -> float)) -> series:Series<'K,float> -> Series<'K,float> Type parameters: 'K	Linearly interpolates an ordered series given a new sequence of keys. Parameters `keys` - Sequence of new keys that forms the index of interpolated results `keyDiff` - A function representing "subtraction" between two keys


            interpolate(keys f series)

Signature: keys:seq<'K> -> (f:('K -> ('K * 'T) option -> ('K * 'T) option -> 'T)) -> series:Series<'K,'T> -> Series<'K,'T>
Type parameters: 'K, 'T

Interpolates an ordered series given a new sequence of keys. The function iterates through each new key, and invokes a function on the current key, the nearest smaller and larger valid observations from the series argument. The function must return a new valid float.

Parameters

keys - Sequence of new keys that forms the index of interpolated results
f - Function to do the interpolating


            interpolateLinear(keys keyDiff series)

Signature: keys:seq<'K> -> (keyDiff:('K -> 'K -> float)) -> series:Series<'K,float> -> Series<'K,float>
Type parameters: 'K

Linearly interpolates an ordered series given a new sequence of keys.

Parameters

keys - Sequence of new keys that forms the index of interpolated results
keyDiff - A function representing "subtraction" between two keys

Series statistics

Functions such as count, mean, kurt etc. return the statistics calculated over all values of a series. The calculation skips over missing values (or nan values), so for example mean returns the average of all present values.

Static members

Static member	Description
`count(series)` Signature: series:Series<'K,'V> -> int Type parameters: 'K, 'V	Returns the number of the values in a series. This excludes missing values and values created from `Double.NaN` etc.
`kurt(series)` Signature: series:Series<'K,float> -> float Type parameters: 'K	Returns the kurtosis of the values in a series. The function skips over missing values and `NaN` values. When there are less than 4 values, the result is NaN.
`max(series)` Signature: series:Series<'K,'V> -> 'V option Type parameters: 'K, 'V	Returns the maximum of the values in a series. The result is an option value. When the series contains no values, the result is `None`.
`maxBy(f series)` Signature: (f:('T -> '?491722)) -> series:Series<'K,'T> -> ('K * 'T) option Type parameters: 'T, '?491722, 'K	Returns the key and value of the greatest element in the series. The result is an optional value. When the series contains no values, the result is `None`.
`mean(series)` Signature: series:Series<'K,float> -> float Type parameters: 'K	Returns the mean of the values in a series. The function skips over missing values and `NaN` values. When there are no available values, the result is NaN.
`median(series)` Signature: series:Series<'K,float> -> float Type parameters: 'K	Returns the median of the elements of the series.
`min(series)` Signature: series:Series<'K,'V> -> 'V option Type parameters: 'K, 'V	Returns the minimum of the values in a series. The result is an option value. When the series contains no values, the result is `None`.
`minBy(f series)` Signature: (f:('T -> '?491726)) -> series:Series<'K,'T> -> ('K * 'T) option Type parameters: 'T, '?491726, 'K	Returns the key and value of the least element in the series. The result is an optional value. When the series contains no values, the result is `None`.
`numSum(series)` Signature: series:Series<'K,^V> -> ^V Type parameters: 'K, ^V	Sum that operates only any appropriate numeric type. When there are no available values, the result is zero of the approriate numeric type.
`skew(series)` Signature: series:Series<'K,float> -> float Type parameters: 'K	Returns the skewness of the values in a series. The function skips over missing values and `NaN` values. When there are less than 3 values, the result is NaN.
`stdDev(series)` Signature: series:Series<'K,float> -> float Type parameters: 'K	Returns the standard deviation of the values in a series. The function skips over missing values and `NaN` values. When there are less than 2 values, the result is NaN.
`sum(series)` Signature: series:Series<'K,float> -> float Type parameters: 'K	Returns the sum of the values in a series. The function skips over missing values and `NaN` values. When there are no available values, the result is NaN.
`variance(series)` Signature: series:Series<'K,float> -> float Type parameters: 'K	Returns the variance of the values in a series. The function skips over missing values and `NaN` values. When there are less than 2 values, the result is NaN.

Deedle

Stats

Remarks

Table of contents

Expanding windows

Static members

Frame statistics

Static members

Moving windows

Static members

Multi-level statistics

Static members

Series interoploation

Static members

Parameters

Parameters

Series statistics

Static members