House_Price_outlier.py

Q1 = []
Q3 = []
Lower_bound = []
Upper_bound = []
Outliers = []


for i in top_features:
    
    # 25th and 75th percentiles
    q1, q3 = np.percentile(train[i],25), np.percentile(train[i],75)
    # Interquartile range
    iqr = q3 - q1
    # Outlier cutoff
    cut_off = 1.5*iqr
    # Lower and Upper bounds
    lower_bound = q1-cut_off
    upper_bound = q3+cut_off
        
    # save outlier indexes
    outlier = [x for x in train.index if train.loc[x,i]<lower_bound or train.loc[x,i]>upper_bound]
    
    # append values for DataFrame
    Q1.append(q1)
    Q3.append(q3)
    Lower_bound.append(lower_bound)
    Upper_bound.append(upper_bound)
    Outliers.append(len(outlier))
    
    try:
        train.drop(outlier,inplace=True,axis=0)
    except:
        continue

df_out = pd.DataFrame({'Column':top_features,'Q1':Q1,'Q3':Q3,'Lower bound':Lower_bound,'Upper_bound':Upper_bound,'No. of outliers':Outliers})    
df_out.sort_values(by='No. of outliers',ascending=False)