memoryfm.weighted_attachment

memoryfm.weighted_attachment(sclog, by='track', freq='W', year=None, alpha=1)

Calculate weighted attachment index.

The Weights are given by log2(1+r), where r is the ratio of scrobbles in a period and maximum number of scrobbles in all periods for the frequency.

Parameters:

sclogScrobbleLog

a memoryfm ScrobbleLog

bytrack, artist, album

The field to use for calculating the weighted Attachment Index.

freqstr, pd.Period

The time frequency to use for the weighted attachment index. For instance, “2W” means a attachment index is calculated bi-weekly i.e. one value for every two weeks. For a full description, see pandas period aliases

yearint, Optional

The year to use for filtering. If None or not passed, no filtering is done.

alphanp.float64

A positive floating point value, representing the order of Attachment Index.

Returns:

pd.Series

Returns a pandas Series containing weighted Attachment Index for each period.