PR by gyandhanee

build.py

@@ -1,26 +1,85 @@
+from __future__ import division
+import numpy as np
+import pandas as pd
+import matplotlib.pyplot as plt
+import operator
+
+
 def get_categorical_variables(df):
-    return []
+    '''Extract the categorical values from a DataFrame'''
+    keys = []
+    values = []
+    dic1 = {}
+    delta_limit = 20
+    #Store the column names sorted by their unique value counts in a list.
+    #Store the sorted values in a list
+    for col in list(df.columns):
+        dic1[col] = len(df[col].unique())
+    dic1 = sorted(dic1.items(), key=operator.itemgetter(1))
+    for k,v in dic1:
+        keys.append(k)
+        values.append(v)
+
+    for i in range(len(values)):
+        next_delta = values[i+1] - values[i] / values[i] #measure the increase in value
+        if next_delta > delta_limit:
+            return df[keys[:i+1]]
 
 
 def get_numerical_variables(df):
-    return []
+
+    return pd.DataFrame._get_numeric_data(df)
 
 
 def get_numerical_variables_percentile(df):
-    pass
+
+    df = get_numerical_variables(df)
+    res_df = df.describe()
+    res_df = res_df.transpose()
+    variable_name = list(res_df.index)
+    res_df['variable name'] = variable_name
+    return res_df
 
 
 def get_categorical_variables_modes(df):
-    pass
+
+    df = get_categorical_variables(df)
+    dic = {'converted':0, 'country':'', 'new_user':0, 'source':''}
+    for col in df.mode().columns:
+        dic[col] = df.mode()[col][0]
+    return pd.DataFrame(dic.items(), columns=['var_name', 'mode'])
 
 
 def get_missing_values_count(df):
-    pass
+
+    ans_df = pd.DataFrame(df[df.isnull()].count(), columns=['missing_value_count'])
+    ans_df.index.name = 'var_name'
+    return ans_df
 
 
 def plot_histogram_with_numerical_values(df):
-    pass
+
+    fig, axes = plt.subplots(2, 2)
+
+    axes[0,0].hist(df[list_of_cols[0]])
+    axes[0,0].set_title(list_of_cols[0])
+
+    axes[0,1].hist(df[list_of_cols[1]])
+    axes[0,1].set_title(list_of_cols[1])
+
+    axes[1,0].hist(df[list_of_cols[2]])
+    axes[1,0].set_title(list_of_cols[2])
+
+    axes[1,1].hist(df[list_of_cols[3]])
+    axes[1,1].set_title(list_of_cols[3])
+
+    plt.tight_layout()
+    plt.show()
 
 
 def plot_facet_box(df):
-    pass
+
+    for col in list_of_cols:
+        plt.boxplot(df[col], 1)
+        plt.title(col)
+        plt.show()