find_outliers.py

# Find potential outliers in values array
# and visualize them on a plot

def is_outlier(value, p25, p75):
    """Check if value is an outlier
    """
    lower = p25 - 1.5 * (p75 - p25)
    upper = p75 + 1.5 * (p75 - p25)
    return value <= lower or value >= upper
 
 
def get_indices_of_outliers(values):
    """Get outlier indices (if any)
    """
    p25 = np.percentile(values, 25)
    p75 = np.percentile(values, 75)
     
    indices_of_outliers = []
    for ind, value in enumerate(values):
        if is_outlier(value, p25, p75):
            indices_of_outliers.append(ind)
    return indices_of_outliers
 

indices_of_outliers = get_indices_of_outliers(dist)
 
fig = plt.figure()
ax = fig.add_subplot(111)
ax.plot(dist, 'b-', label='distances')
ax.plot(
    indices_of_outliers,
    values[indices_of_outliers],
    'ro',
    markersize = 7,
    label='outliers')
ax.legend(loc='best')