Commit Live PR

q01_load_data/build.py

@@ -1,5 +1,8 @@
 import pandas as pd
 
-
+path = 'data/excel-comp-data.xlsx'
 def q01_load_data(path):
-    "write your solution here"
+    df = pd.read_excel(path)
+    df['state'] = df['state'].str.lower()
+    df['total'] = df['Jan'] + df['Feb'] + df['Mar']
+    return df

q02_append_row/build.py

@@ -1,12 +1,17 @@
 import pandas as pd
 import sys, os
-#sys.path.append(os.path.join(os.path.dirname(os.curdir)))
-from greyatomlib.pandas_guided_project.q01_load_data.build import q01_load_data
-
-
+sys.path.append(os.path.join(os.path.dirname(os.curdir)))
+#from greyatomlib.pandas_guided_project.q01_load_data.build import q01_load_data
+from q01_load_data.build import q01_load_data
+path = 'data/excel-comp-data.xlsx'
 def q02_append_row(path):
-    "write your solution here"
-
-
-
-
+    df = q01_load_data(path)
+    df1 = df[['Jan', 'Feb', 'Mar', 'total']].sum(axis=0)
+    df1 = pd.DataFrame(df1)
+    df2 = pd.DataFrame(columns=['Jan', 'Feb', 'Mar', 'total'])
+    df2['Jan'] = df1.loc['Jan']
+    df2['Feb'] = df1.loc['Feb']
+    df2['Mar'] = df1.loc['Mar']
+    df2['total'] = df1.loc['total']
+    df3 = df.append(df2, ignore_index=True)
+    return df3

q03_scrape_clean/build.py

@@ -3,7 +3,27 @@
 import requests
 sys.path.append(os.path.join(os.path.dirname(os.curdir)))
 
-
+path = 'https://en.wikipedia.org/wiki/List_of_U.S._state_abbreviations'
 def q03_scrape_clean(url):
-    "write your solution here"
+    page = requests.get(url)
+    df = pd.read_html(page.text)
+    df1 = df[0].iloc[11:,0]
+    df2 = df[0].iloc[11:,1]
+    df3 = df[0].iloc[11:,2]
+    df4 = df[0].iloc[11:,3]
 
+    df5 = df[0].iloc[11:,4]
+    df6 = df[0].iloc[11:,5]
+    df7 = df[0].iloc[11:,6]
+    df8 = df[0].iloc[11:,7]
+    df9 = df[0].iloc[11:,8]
+    df10 = df[0].iloc[11:,9]
+    ans = pd.concat([df1,df2,df3,df4,df5,df6,df7,df8,df9,df10], axis=1)
+    ans.rename(mapper={0:'Name', 1:'Status', 2:'ISO', 3:'ANSI0', 4:'ANSI1', 5:'USPS', 6:'USCG', 7:'GPO', 8:'AP', 9:'Other Abbrevations'}, inplace=True, axis=1)
+    ans.drop(ans.index[0], axis=0, inplace=True)
+    ans['ex1'] = 0
+    ans['ex2'] = 0
+    ans['ex3'] = 0
+    ans['ex4'] = 0
+    ans['ex5'] = 0
+    return ans

q04_mapping/build.py

@@ -1,10 +1,28 @@
 import pandas as pd
 import sys, os
 import numpy as np
-#sys.path.append(os.path.join(os.path.dirname(os.curdir)))
-from greyatomlib.pandas_guided_project.q02_append_row.build import q02_append_row
+sys.path.append(os.path.join(os.path.dirname(os.curdir)))
+from q02_append_row.build import q02_append_row
+path1 = 'data/excel-comp-data.xlsx'
+path2 = 'data/scraped.csv'
+
 def q04_mapping(path1,path2):
-    "write your solution here"
+    df1 = q02_append_row(path1)
+    df1['abbr'] = np.nan
+    df2 = pd.read_csv(path2)
+    ab = df2.iloc[:,7]
+    name = df2['United States of America']
+    d = {}
+    for i in range(0,ab.shape[0]):
+        d[name[i].lower()] = ab[i]
 
+    for i in range(0,df1.shape[0]):
+        if df1.iloc[i,:]['state'] in d.keys():
+            df1.iloc[i,-1] = d[df1.iloc[i,:]['state']]
 
+    df2 = df1.iloc[:,0:5]
+    df2['total'] = df1['total']
+    df2['abbr'] = df1['abbr']
+    return df2
 
+q04_mapping(path1,path2)

q05_replace_missing_values/build.py

@@ -2,12 +2,16 @@
 import numpy as np
 import sys
 import os
-#sys.path.append(os.path.join(os.path.dirname(os.curdir)))
-from greyatomlib.pandas_guided_project.q04_mapping.build import q04_mapping
+sys.path.append(os.path.join(os.path.dirname(os.curdir)))
+from q04_mapping.build import q04_mapping
 
 path1 = 'data/excel-comp-data.xlsx'
 path2 = 'data/scraped.csv'
 def q05_replace_missing_values(path1,path2):
+    df = q04_mapping(path1,path2)
+    df.iloc[6,6] = 'MS'
+    df.iloc[10,6] = 'TN'
+    return df
 
 
-#print(q05_replace_missing_values(path1,path2).shape)
+#q05_replace_missing_values(path1,path2)

q06_sub_total/build.py

@@ -3,14 +3,16 @@
 from sklearn.model_selection import train_test_split
 import sys
 import os
-#sys.path.append(os.path.join(os.path.dirname(os.curdir)))
-from greyatomlib.pandas_guided_project.q05_replace_missing_values.build import q05_replace_missing_values
+sys.path.append(os.path.join(os.path.dirname(os.curdir)))
+from q05_replace_missing_values.build import q05_replace_missing_values
 
 path1 = 'data/excel-comp-data.xlsx'
 path2 = 'data/scraped.csv'
 
 def q06_sub_total(path1,path2):
-    "write your solution here"
-
-
+    df = q05_replace_missing_values(path1,path2)
+    #print(df)
+    df1 = df.groupby('abbr').sum()
+    return df1
 
+#q06_sub_total(path1,path2)

q07_symbols/build.py

@@ -4,15 +4,19 @@
 import sys
 import os
 sys.path.append(os.path.join(os.path.dirname(os.curdir)))
-from greyatomlib.pandas_guided_project.q06_sub_total.build import q06_sub_total
+from q06_sub_total.build import q06_sub_total
 
 path1 = 'data/excel-comp-data.xlsx'
 path2 = 'data/scraped.csv'
 
 def q07_symbols(path1,path2):
-    "write your solution here"
+    df = q06_sub_total(path1,path2)
+    df1 = df.applymap(lambda x: str('$')+str(x))
+    df2 = df1.applymap(lambda x: str(x[0:4])+str(',')+str(x[4:]))
+    return df2
 
+#q07_symbols(path1,path2)
 
 
-#print(q07_symbols(path1,path2))
 
+#print(q07_symbols(path1,path2))

q08_append_subtotals/build.py

@@ -1,15 +1,38 @@
 import pandas as pd
 import numpy as np
 import sys,os
-#sys.path.append(os.path.join(os.path.dirname(os.curdir)))
-from greyatomlib.pandas_guided_project.q06_sub_total.build import q06_sub_total
-from greyatomlib.pandas_guided_project.q07_symbols.build import q07_symbols
+sys.path.append(os.path.join(os.path.dirname(os.curdir)))
+from q06_sub_total.build import q06_sub_total
+from q07_symbols.build import q07_symbols
 
 
 path1 = 'data/excel-comp-data.xlsx'
 path2 = 'data/scraped.csv'
 
-def q08_append_subtotals(path1,path2):
-    "write your solution here"
+def comma(x):
+    if len(str(x)) == 6:
+        x = str(x)
+        return str('$')+str(x[0:3])+str(',')+str(x[3:])
+
+    elif len(str(x)) == 5:
+        x = str(x)
+        return str('$')+str(x[0:2])+str(',')+str(x[2:])
 
+    elif len(str(x)) == 7:
+        x = str(x)
+        return str('$')+str(x[0])+str(',')+str(x[1:4])+str(',')+str(x[4:])
+
+def q08_append_subtotals(path1,path2):
+    df = q06_sub_total(path1,path2)
+    jan = df['Jan'].sum()
+    feb = df['Feb'].sum()
+    mar = df['Mar'].sum()
+    t = df['total'].sum()
+    a = df['account'].sum()
+    temp = pd.DataFrame({'Jan':[jan],'Feb':[feb], 'Mar':[mar], 'account':[a], 'total':[t]})
+    df2 = df.append(temp,ignore_index=True)
+    dfy = df2.applymap(comma)
+    dfy.drop('account', axis=1, inplace=True)
+    return dfy
 
+#q08_append_subtotals(path1,path2)

q09_pie_chart_jan/build.py

@@ -3,13 +3,12 @@
 from sklearn.model_selection import train_test_split
 import sys,os
 sys.path.append(os.path.join(os.path.dirname(os.curdir)))
-from greyatomlib.pandas_guided_project.q06_sub_total.build import q06_sub_total
+from q06_sub_total.build import q06_sub_total
+import matplotlib
+matplotlib.use('Agg')
 import matplotlib.pyplot as plt
-plt.switch_backend('agg')
 
 def q09_pie_chart_jan(path1,path2):
-
-    "write your solution here"
-
-
-
+    df = q06_sub_total(path1,path2)
+    df.plot(x='Jan', kind='pie', subplots='True')
+    plt.show()

q10_total/build.py

@@ -3,12 +3,12 @@
 from sklearn.model_selection import train_test_split
 import sys,os
 sys.path.append(os.path.join(os.path.dirname(os.curdir)))
-from greyatomlib.pandas_guided_project.q06_sub_total.build import q06_sub_total
+from q06_sub_total.build import q06_sub_total
+import matplotlib
+matplotlib.use('Agg')
 import matplotlib.pyplot as plt
-plt.switch_backend('agg')
 
 def q10_total(path1,path2):
-
-    "write your solution here"
-
-
+    df = q06_sub_total(path1,path2)
+    df.plot(x='Jan', kind='pie', subplots='True')
+    plt.show()