weightedMovingAverage.py

def weightedMovingAverage(period, data):

    # --- Weighted Moving Average
    # data: array, time series data e.g. daily close prices
    # period: integer, number of periods from time series array to include in calculation

    import numpy as np

    # --- get first non nan index
    for i in range(len(data)):

        if np.isnan(data[i]) == False:

            firstNonNan = i
            break

    # --- get last non nan index
    for i in reversed(range(len(data))):

        if np.isnan(data[i]) == False:

            lastNonNan = i
            break

    # --- based on period value calculate the 'weighting factor' or denominator
    d = period * (period / 2 + 0.5)

    # --- generate weights list
    weights = np.zeros(period)
    for i in range(period):
        weights[i] = (i + 1) / d

    # --- calculate WMA
    a = np.zeros(period)
    out = np.zeros(len(data))

    for i in range(len(data)):

        if i < firstNonNan + period - 1:
            out[i] = np.nan

        elif i > lastNonNan:
            out[i] = np.nan

        else:
            for m in range(period):
                a[(period - 1) - m] = data[
                    i - m,
                ]
            out[i] = np.sum(a * weights)

    return out