EDA PLOTS:-

~~~~~~~~~~~~~~~~~~~~~~~~~~~
if target_type1 == 'Continuous':
df = first_data[feat_imp1]
date = ['date','Date','DATE']
col = df.columns
def rem_date(x):
for dates in date:
for cols in col:
if dates == cols :
df_new = x.drop(dates,axis = 1)
print("date column removed")
return df_new
else:
print("date column not found")
return df

df = rem_date(df)
print(df.head())

symbols = ["?","#","%","&","*","+","@","_","-"]
def remove_spchar(x):
for column in x.values:
for symbol in symbols:
if symbol not in column:
print("Special Character Not detected")
return x
else:
print("Special Character Detected and will be Removed: ")
x2 = x.replace(symbol, 0)
return x2

df = remove_spchar(df)
df.head()
def validate_null_values(x):
check = x.isnull().sum()
val = check.values
val2 = 0
for col in x.columns:
for ele in range(0, len(val)):
val2 = val2 + val[ele]
if val2 == 0:
print("There is no null values:")
return x
else:
data_n = x.fillna(x[col].mean())
print("There is null values and will be removed:")
return data_n

df = validate_null_values(df)
df.head()
ids_ = ['Id','ID','id','encounter_id']
cols = df.columns
def rem_id(x):
for id_ in ids_:
for col in cols:
if col == id_:
new_df = x.drop(id_,axis = 1)
print("id column removed")
 return new_df
else:
print("id column not found")
return x

df = rem_id(df)
df.head()
X = df.drop(target_var1,axis = 1)
Y = df[target_var1]
from sklearn.preprocessing import LabelEncoder
le = LabelEncoder()
X = X.apply(lambda col: le.fit_transform(col.astype(str)), axis = 0,
result_type = 'expand')
X.head()

from sklearn.model_selection import train_test_split

X_train,X_test,y_train,y_test =
train_test_split(X,Y,test_size=.20,random_state=0)
from sklearn.feature_selection import f_classif, f_regression
univariate = f_regression(X_train.fillna(0), y_train)
univariate = pd.Series(univariate[1])
univariate.index = X_train.columns
univ = pd.DataFrame(univariate).rename_axis('Feature').reset_index()
univ.rename(columns={0: 'ANOVA'}, inplace=True)
univ_o = univ.sort_values(by='ANOVA',ascending=False)
thresh = round(univ_o.shape[0] *0.7)
univ_o = univ_o.iloc[:thresh,:-1]
univ_o = list(univ_o['Feature'])
from sklearn.ensemble import RandomForestRegressor
model = RandomForestRegressor(max_features=None).fit(X_train, y_train)
dt_fi = np.array(model.feature_importances_)
feature_names = np.array(X_train.columns)
data1={'Feature':feature_names,'Model':dt_fi}
dt_df = pd.DataFrame(data1)
dt_o = dt_df.sort_values(by='Model',ascending=False)
dt_o = dt_o.iloc[:thresh,:-1]
dt_o = list(dt_o['Feature'])
print(dt_o)
pred = model.predict(X_test)
y_test = y_test.values.flatten()

# actual vs predicted plot

import matplotlib.pyplot as plt
true_value = y_test
predicted_value = pred
plt.figure(figsize=(8, 5))
plt.scatter(true_value, predicted_value)
plt.yscale('log')
plt.xscale('log')

p1 = max(max(predicted_value), max(true_value))
p2 = min(min(predicted_value), min(true_value))
plt.plot([p1, p2], [p1, p2], 'b-')
plt.xlabel('True Values', fontsize=15)
plt.ylabel('Predictions', fontsize=15)
plt.title('Actual Vs predictions')
# plt.axis('equal')
rand = random.randint(00000000, 99999999)
name = 'modelgraph' + str(rand)
 buyerid = request.session['buyerid']
modelgraph= 'media/files/expert_serv/' + str(buyerid) + '/' + str(
name) + '.png'
print(modelgraph)
plt.savefig('media/files/expert_serv/' + str(buyerid) +'/' + str(name) +
'.png', dpi=300)

# feature importance plot

imp = model.feature_importances_
imp_sort = np.sort(imp)
# X.columns
col = ['value']
fra = pd.DataFrame(imp_sort,X.columns,columns=col)
sort_fra = fra.sort_values(by=['value'],ascending=False)[0:5]
nameimp = sort_fra.index[0:6]
sort_mean_abs = sort_fra.values[0:6]
plt.xticks(rotation = 90)
bar = sort_fra.plot(kind='bar',figsize=(10,5) )
# bar.plot(figsize=(15,8));
rand = random.randint(00000000, 99999999)
name = 'feature' + str(rand)
buyerid = request.session['buyerid']
feature= 'media/files/expert_serv/' + str(buyerid) + '/' + str(
name) + '.png'
print(feature)
plt.savefig('media/files/expert_serv/' + str(buyerid) +'/' + str(name) +
'.png', dpi=300)
# correlation plot
import seaborn as sns
import matplotlib.pyplot as plt
top_5_feature = df[dt_o].iloc[:, 0:6]
top_5_feature.head()
plt.figure(figsize=(15,5))
sns.heatmap(top_5_feature.corr(),annot = True)
rand = random.randint(00000000, 99999999)
name = 'matrix' + str(rand)
buyerid = request.session['buyerid']
matrix2= 'media/files/expert_serv/' + str(buyerid) + '/' + str(
name) + '.png'
print(matrix2)
plt.savefig('media/files/expert_serv/' + str(buyerid) +'/' + str(name) +
'.png', dpi=300)

numerics = ['int16', 'int32','int64', 'float16', 'float32', 'float64']

numerical_features =
list(top_5_feature.select_dtypes(include=numerics).columns)
data_num = df[numerical_features]

obj = ['object']
obj_features = list(top_5_feature.select_dtypes(include=obj).columns)
data_obj = df[obj_features]
num_col_df = df[numerical_features]
obj_col_df = df[obj_features]
# data analysis plots
data_num[target_var1] = Y
x_range = list(range(len(num_col_df.columns)))
print(x_range)
for i in x_range:
try:
 if i ==0:
plt.figure(figsize=(10,5))
sns.scatterplot(x = data_num.columns[-2],y = data_num.columns[-
1],data=data_num)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis1_' + str(rand)
buyerid = request.session['buyerid']
dataanalysis1= 'media/files/expert_serv/' + str(buyerid) + '/'
+ str(
name) + '.png'
print(dataanalysis1)
plt.savefig('media/files/expert_serv/' + str(buyerid) +'/' +
str(name) + '.png', dpi=300)
except:
pass
try:
if i==1:
plt.figure(figsize=(10,5))
sns.lineplot(x = data_num.columns[-3],y = data_num.columns[-
1],data=data_num)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis2_' + str(rand)
buyerid = request.session['buyerid']
dataanalysis2= 'media/files/expert_serv/' + str(buyerid) + '/'
+ str(
name) + '.png'
print(dataanalysis2)
plt.savefig('media/files/expert_serv/' + str(buyerid) +'/' +
str(name) + '.png', dpi=300)
except:
pass

# try:
# if i==2:
# plt.figure(figsize=(10,5))
# sns.scatterplot(x = data_num.columns[-4],y =
data_num.columns[-1],data=data_num)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis3_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis3=location)
# except:
# pass

# try:
# if i==3:
# plt.figure(figsize=(10,5))
# sns.scatterplot(x = data_num.columns[-5],y =
data_num.columns[-1],data=data_num)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis4_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis4=location)
# except:
# pass

# try:
# if i==4:
# plt.figure(figsize=(10,5))
# sns.scatterplot(x = data_num.columns[-6],y =
data_num.columns[-1],data=data_num)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis4_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis4=location)
# except:
# pass
count_col = obj_col_df.columns
data_obj[target_var1] = Y
x_range = list(range(len(count_col)))
print(x_range)

for i in x_range:
try:
if i ==0:
plt.figure(figsize=(10,5))
plt.xticks(rotation=90)
sns.barplot(x = data_obj.columns[-2],y = data_obj.columns[-
1],data=data_obj)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis1_' + str(rand)
buyerid = request.session['buyerid']
dataanalysis1_2= 'media/files/expert_serv/' + str(buyerid)
+ '/' + str(
name) + '.png'
print(dataanalysis1_2)
plt.savefig('media/files/expert_serv/' + str(buyerid) +'/'
+ str(name) + '.png', dpi=300)
else:
pass
except:
pass
try:
if i==1:
plt.figure(figsize=(10,5))
plt.xticks(rotation=90)
sns.countplot(x = data_obj.columns[-3],data=data_obj)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis2_2' + str(rand)
buyerid = request.session['buyerid']
 dataanalysis2_2= 'media/files/expert_serv/' + str(buyerid)
+ '/' + str(
name) + '.png'
print(dataanalysis2_2)
plt.savefig('media/files/expert_serv/' + str(buyerid) +'/'
+ str(name) + '.png', dpi=300)
else:
pass
except:
pass

try:
if i ==2:
plt.figure(figsize=(10,5))
plt.xticks(rotation=90)
sns.violinplot(x = data_obj.columns[-4],y =
data_obj.columns[-1],data=data_obj)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis2_3' + str(rand)
buyerid = request.session['buyerid']
dataanalysis2_3= 'media/files/expert_serv/' + str(buyerid)
+ '/' + str(
name) + '.png'
print(dataanalysis2_3)
plt.savefig('media/files/expert_serv/' + str(buyerid) +'/'
+ str(name) + '.png', dpi=300)
else:
pass
except:
pass

# try:
# if i==3:
# plt.figure(figsize=(10,5))
# plt.xticks(rotation=90)
# sns.countplot(x = data_obj.columns[-5],data=data_obj)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis2_4_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis2_4=location)
# except:
# pass

# try:
# if i==4:
# plt.figure(figsize=(10,5))
# plt.xticks(rotation=90)
# sns.boxplot(x = data_num.columns[-6],y =
data_obj.columns[-1],data=data_obj)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis2_5_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis2_5=location)
# except:
# pass
temp = expert_eda()
temp.expertservice_id = expertservice_id
temp.buyer_id = buyerid
temp.dataanalysis1 = location_1
temp.dataanalysis2 = location_2
temp.dataanalysis2_1= location_2_1
temp.dataanalysis2_2 = location_2_2
temp.dataanalysis2_3 = location_2_3
temp.matrix2 = matrix2
temp.feature = feature
temp.modelgraph = modelgraph
temp.target = target_var1
# temp.algo_type = 'clf'
temp.save()

elif target_type1 == 'Categorical':

df = first_data[feat_imp1]
date = ['date','Date','DATE']
col = df.columns
def rem_date(x):
for dates in date:
for cols in col:
if dates == cols :
df_new = x.drop(dates,axis = 1)
print("date column removed")
return df_new
else:
print("date column not found")
return df

df = rem_date(df)
print(df.head())

symbols = ["?","#","%","&","*","+","@","_","-"]
def remove_spchar(x):
for column in x.values:
for symbol in symbols:
if symbol not in column:
print("Special Character Not detected")
return x
else:
print("Special Character Detected and will be Removed: ")
x2 = x.replace(symbol, 0)
return x2

df = remove_spchar(df)
df.head()
def validate_null_values(x):
check = x.isnull().sum()
val = check.values
val2 = 0
 for col in x.columns:
for ele in range(0, len(val)):
val2 = val2 + val[ele]
if val2 == 0:
print("There is no null values:")
return x
else:
data_n = x.fillna(x[col].mean())
print("There is null values and will be removed:")
return data_n

df = validate_null_values(df)
df.head()
ids_ = ['Id','ID','id','encounter_id']
cols = df.columns
def rem_id(x):
for id_ in ids_:
for col in cols:
if col == id_:
new_df = x.drop(id_,axis = 1)
print("id column removed")
return new_df
else:
print("id column not found")
return x

df = rem_id(df)
df.head()
X = df.drop(target_var1,axis = 1)
Y = df[target_var1]
from sklearn.preprocessing import LabelEncoder
le = LabelEncoder()
X = X.apply(lambda col: le.fit_transform(col.astype(str)), axis = 0,
result_type = 'expand')
X.head()

from sklearn.model_selection import train_test_split

X_train,X_test,y_train,y_test =
train_test_split(X,Y,test_size=.20,random_state=0)
from sklearn.feature_selection import f_classif, f_regression
univariate = f_classif(X_train.fillna(0), y_train)
univariate = pd.Series(univariate[1])
univariate.index = X_train.columns
univ = pd.DataFrame(univariate).rename_axis('Feature').reset_index()
univ.rename(columns={0: 'ANOVA'}, inplace=True)
univ_o = univ.sort_values(by='ANOVA',ascending=False)
thresh = round(univ_o.shape[0] *0.7)
univ_o = univ_o.iloc[:thresh,:-1]
univ_o = list(univ_o['Feature'])
from sklearn.ensemble import RandomForestClassifier
model = RandomForestClassifier(max_features=None).fit(X_train, y_train)
dt_fi = np.array(model.feature_importances_)
feature_names = np.array(X_train.columns)
data1={'Feature':feature_names,'Model':dt_fi}
dt_df = pd.DataFrame(data1)
dt_o = dt_df.sort_values(by='Model',ascending=False)
dt_o = dt_o.iloc[:thresh,:-1]
dt_o = list(dt_o['Feature'])
print(dt_o)
 pred = model.predict(X_test)
y_test = y_test.values.flatten()

# actual vs predicted plot

import matplotlib.pyplot as plt
true_value = y_test
predicted_value = pred
plt.figure(figsize=(8, 5))
plt.scatter(true_value, predicted_value)
plt.yscale('log')
plt.xscale('log')

p1 = max(max(predicted_value), max(true_value))
p2 = min(min(predicted_value), min(true_value))
plt.plot([p1, p2], [p1, p2], 'b-')
plt.xlabel('True Values', fontsize=15)
plt.ylabel('Predictions', fontsize=15)
plt.title('Actual Vs predictions')
# plt.axis('equal')
rand = random.randint(00000000, 99999999)
name = 'modelgraph_new' + str(rand)
buyerid = request.session['buyerid']
modelgraph = 'media/files/expert_serv/' + str(buyerid) + '/' + str(name) +
'.png'
print(modelgraph)
plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' + str(name) +
'.png', dpi=300)
# info = expert_eda.objects.filter(id=esp).all()
# info.update(modelgraph=location)

# feature importance plot

imp = model.feature_importances_
imp_sort = np.sort(imp)
# X.columns
col = ['value']
fra = pd.DataFrame(imp_sort,X.columns,columns=col)
sort_fra = fra.sort_values(by=['value'],ascending=False)[0:5]
nameimp = sort_fra.index[0:6]
sort_mean_abs = sort_fra.values[0:6]
plt.xticks(rotation = 90)
bar = sort_fra.plot(kind='bar',figsize=(10,5) )
# bar.plot(figsize=(15,8));
name = 'feature_new' + str(rand)
buyerid = request.session['buyerid']
feature = 'media/files/expert_serv/' + str(buyerid) + '/' + str(name) +
'.png'
print(feature)
plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' + str(name) +
'.png', dpi=300)
# info = expert_eda.objects.filter(id=esp).all()
# info.update(feature=location)
# correlation plot
import seaborn as sns
import matplotlib.pyplot as plt
top_5_feature = df[dt_o].iloc[:, 0:6]
top_5_feature.head()
plt.figure(figsize=(15,5))
sns.heatmap(top_5_feature.corr(),annot = True)
rand = random.randint(00000000, 99999999)
 name = 'matrix_new' + str(rand)
buyerid = request.session['buyerid']
matrix2 = 'media/files/expert_serv/' + str(buyerid) + '/' + str(name) +
'.png'
print(matrix2)
plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' + str(name) +
'.png', dpi=300)
# info = expert_eda.objects.filter(id=esp).all()
# info.update(matrix=location)

numerics = ['int16', 'int32','int64', 'float16', 'float32', 'float64']

numerical_features =
list(top_5_feature.select_dtypes(include=numerics).columns)
data_num = df[numerical_features]

obj = ['object']
obj_features = list(top_5_feature.select_dtypes(include=obj).columns)
data_obj = df[obj_features]
num_col_df = df[numerical_features]
obj_col_df = df[obj_features]
# data analysis plots
data_num[target_var1] = Y
x_range = list(range(len(num_col_df.columns)))
print(x_range)
for i in x_range:
try:
if i ==0:
plt.figure(figsize=(10,5))
sns.scatterplot(x = data_num.columns[-2],y = data_num.columns[-
1],data=data_num)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis1_new' + str(rand)
buyerid = request.session['buyerid']
location_1 = 'media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png'
print(location_1)
plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png', dpi=300)
# info = expert_eda.objects.filter(id=esp).all()
# info.update(dataanalysis1=location)
except:
pass
try:
if i==1:
plt.figure(figsize=(10,5))
sns.lineplot(x = data_num.columns[-3],y = data_num.columns[-
1],data=data_num)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis2_new' + str(rand)
buyerid = request.session['buyerid']
location_2 = 'media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png'
print(location_2)
plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png', dpi=300)
# info = expert_eda.objects.filter(id=esp).all()
# info.update(dataanalysis2=location)
except:
pass
 # try:
# if i==2:
# plt.figure(figsize=(10,5))
# sns.scatterplot(x = data_num.columns[-4],y =
data_num.columns[-1],data=data_num)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis3_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis3=location)
# except:
# pass

# try:
# if i==3:
# plt.figure(figsize=(10,5))
# sns.scatterplot(x = data_num.columns[-5],y =
data_num.columns[-1],data=data_num)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis4_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis4=location)
# except:
# pass

# try:
# if i==4:
# plt.figure(figsize=(10,5))
# sns.scatterplot(x = data_num.columns[-6],y =
data_num.columns[-1],data=data_num)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis4_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) + '/' +
str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis4=location)
# except:
# pass
count_col = obj_col_df.columns
data_obj[target_var1] = Y
x_range = list(range(len(count_col)))
print(x_range)
 for i in x_range:
try:
if i ==0:
plt.figure(figsize=(10,5))
plt.xticks(rotation=90)
sns.barplot(x = data_obj.columns[-2],y = data_obj.columns[-
1],data=data_obj)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis2_1_new' + str(rand)
buyerid = request.session['buyerid']
location_2_1 = 'media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png'
print(location_2_1)
plt.savefig('media/files/expert_serv/' + str(buyerid) + '/'
+ str(name) + '.png', dpi=300)
# info = expert_eda.objects.filter(id=esp).all()
# info.update(dataanalysis2_1=location)
except:
pass
try:
if i==1:
plt.figure(figsize=(10,5))
plt.xticks(rotation=90)
sns.countplot(x = data_obj.columns[-3],data=data_obj)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis2_2_new' + str(rand)
buyerid = request.session['buyerid']
location_2_2 = 'media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png'
print(location_2_2)
plt.savefig('media/files/expert_serv/' + str(buyerid) + '/'
+ str(name) + '.png', dpi=300)
# info = expert_eda.objects.filter(id=esp).all()
# info.update(dataanalysis2_2=location)
except:
pass

try:
if i ==2:
plt.figure(figsize=(10,5))
plt.xticks(rotation=90)
sns.violinplot(x = data_obj.columns[-4],y =
data_obj.columns[-1],data=data_obj)
rand = random.randint(00000000, 99999999)
name = 'dataanalysis2_3_new' + str(rand)
buyerid = request.session['buyerid']
location_2_3 = 'media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png'
print(location_2_3)
plt.savefig('media/files/expert_serv/' + str(buyerid) + '/'
+ str(name) + '.png', dpi=300)
# info = expert_eda.objects.filter(id=esp).all()
# info.update(dataanalysis2_3=location)
except:
pass

# try:
# if i==3:
# plt.figure(figsize=(10,5))
 # plt.xticks(rotation=90)
# sns.countplot(x = data_obj.columns[-5],data=data_obj)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis2_4_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis2_4=location)
# except:
# pass

# try:
# if i==4:
# plt.figure(figsize=(10,5))
# plt.xticks(rotation=90)
# sns.boxplot(x = data_num.columns[-6],y =
data_obj.columns[-1],data=data_obj)
# rand = random.randint(00000000, 99999999)
# name = 'dataanalysis2_5_' + str(rand)
# buyerid = request.session['buyerid']
# location = 'media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png'
# print(location)
# plt.savefig('media/files/expert_serv/' + str(buyerid) +
'/' + str(name) + '.png', dpi=300)
# info = expertservice.objects.filter(id=esp).all()
# info.update(dataanalysis2_5=location)
# except:
# pass
temp = expert_eda()
temp.expertservice_id = expertservice_id
temp.buyer_id = buyerid
temp.dataanalysis1 = location_1
temp.dataanalysis2 = location_2
temp.dataanalysis2_1= location_2_1
temp.dataanalysis2_2 = location_2_2
temp.dataanalysis2_3 = location_2_3
temp.matrix2 = matrix2
temp.feature = feature
temp.modelgraph = modelgraph
temp.target = target_var1
# temp.algo_type = 'clf'
temp.save()
else:
pass