def make_prediction():
model = joblib.load('model.pkl')
if request.method == 'POST':
entered_li = []
month = request.form['Month']
#day = int(request.form['Day'])
promo = int(request.form['Promo'])
#promo2 = int(request.form['Promo2'])
stateH = int(request.form['StateH'])
schoolH = int(request.form['SchoolH'])
assortment = int(request.form['Assortment'])
storeType = int(request.form['StoreType'])
store = int(request.form['store'])
date_entry = month
year, month, day = map(int, date_entry.split('-'))
# one-hot encode categorical variables
stateH_encode = onehotCategorical(stateH, 4)
assortment_encode = onehotCategorical(assortment, 3)
storeType_encode = onehotCategorical(storeType, 4)
store_encode = onehotCategorical(store, 1115, store=1)
comp_dist = 5458.1
entered_li.extend(store_encode)
entered_li.extend(storeType_encode)
entered_li.extend(assortment_encode)
entered_li.extend(stateH_encode)
entered_li.extend([comp_dist])
#entered_li.extend([promo2])
entered_li.extend([promo])
entered_li.extend([day])
entered_li.extend([month])
entered_li.extend([schoolH])
data = [[
store, 1270, promo, schoolH, storeType, assortment, stateH, 6, day,
month, year, 50, 132, 0, 0
]]
df = pd.DataFrame(data,
columns=[
'Store', 'CompetitionDistance', 'Promo',
'SchoolHoliday', 'StoreType', 'Assortment',
'StateHoliday', 'DayOfWeek', 'Month', 'Day',
'Year', 'WeekOfYear', 'CompetitionOpen',
'PromoOpen', 'IsPromoMonth'
])
#data = [[1,1270,1,0,3,1,0,6,9,15,2019,37,132,0,0]]
#df = pd.DataFrame(data,columns=['Store','CompetitionDistance','Promo','SchoolHoliday','StoreType','Assortment','StateHoliday','DayOfWeek','Month','Day','Year','WeekOfYear','CompetitionOpen','PromoOpen','IsPromoMonth'])
prediction = model.predict(xgb.DMatrix(df))
prediction = np.expm1(prediction)
#prediction = model.predict(entered_li.values.reshape(1, -1))
label = "$" + str(np.squeeze(prediction.round(2)))
return render_template('index.html', label=label)
def make_prediction():
if request.method == 'POST':
entered_li = []
# ========== Part 2.3 ==========
# YOUR CODE START HERE
data_month = request.form['month']
data_promo = request.form['promo']
data_stateHoliday = request.form['state_holiday']
data_assortment = request.form['assortment']
data_store = request.form['store']
data_dayOfWeek = request.form['day_of_the_week']
data_promo2 = request.form['promo2']
data_schoolHoliday = request.form['school_holiday']
data_storeType = request.form['store_type']
# get request values
# one-hot encode categorical variables (only for categorical variables)
data1 = onehotCategorical(int(data_store), 1112, 0)
data2 = onehotCategorical(int(data_storeType), 4, 1)
data3 = onehotCategorical(int(data_assortment), 3, 1)
data4 = onehotCategorical(int(data_stateHoliday), 4, 1)
# manually specify competition distance
comp_dist = 5458.1
# build 1 observation for prediction
temp1 = np.append(data1, data2)
temp1 = np.append(temp1, data3)
temp1 = np.append(temp1, data4)
temp2 = np.array([
comp_dist,
int(data_promo2),
int(data_promo),
int(data_dayOfWeek),
int(data_month),
int(data_schoolHoliday)
])
entered_li = np.append(temp1, temp2)
# ========== End of Part 2.3 ==========
# make prediction
prediction = model.predict(np.array(entered_li).reshape(1, -1))
label = str(np.squeeze(prediction.round(2)))
return render_template('index.html', label=label)
def make_prediction():
if request.method=='POST':
entered_li = []
# YOUR CODE FOR PART 2.2
# get request values
month = int(request.form['Month'])
day = int(request.form['Day'])
promo = int(request.form['Promo'])
promo2 = int(request.form['Promo2'])
stateH = int(request.form['StateH'])
schoolH = int(request.form['SchoolH'])
assortment = int(request.form['Assortment'])
storeType = int(request.form['StoreType'])
store = int(request.form['store'])
# one-hot encode categorical variables
stateH_encode = onehotCategorical(stateH, 4)
assortment_encode = onehotCategorical(assortment, 3)
storeType_encode = onehotCategorical(storeType, 4)
store_encode = onehotCategorical(store, 1115, store=1)
# manually specify competition distance
comp_dist = 5458.1
# engineer 1 observation for prediction
# YOUR CODE START HERE
entered_li.extend(store_encode)
entered_li.extend(storeType_encode)
entered_li.extend(assortment_encode)
entered_li.extend(stateH_encode)
entered_li.extend([comp_dist])
entered_li.extend([promo2])
entered_li.extend([promo])
entered_li.extend([day])
entered_li.extend([month])
entered_li.extend([schoolH])
#pkl_file = open('lr.pkl', 'rb')
#model = pickle.load(pkl_file)
model = joblib.load('lr.pkl')
prediction = model.predict(np.array(entered_li).reshape(1, -1))
#prediction = model.predict(entered_li.values.reshape(1, -1))
label = str(np.squeeze(prediction.round(2)))
return render_template('index.html', label=label)
def make_prediction():
if request.method == 'POST':
entered_li = []
# ========== Part 2.3 ==========
# YOUR CODE START HERE
# get request values
store = int(request.form['store'])
month = int(request.form['month'])
promo = int(request.form['promo'])
state_holiday = int(request.form['state_holiday'])
assortment = int(request.form['assortment'])
day_of_the_week = int(request.form['day_of_the_week'])
promo2 = int(request.form['promo2'])
school_holiday = int(request.form['school_holiday'])
store_type = int(request.form['store_type'])
# one-hot encode categorical variables
onehotStore = onehotCategorical(store, 945, 1)
onehotStore_type = onehotCategorical(store_type, 4)
onehotAssortment = onehotCategorical(assortment, 3)
onehotState_holiday = onehotCategorical(state_holiday, 4)
# manually specify competition distance
comp_dist = 5458.1
# build 1 observation for prediction
entered_li.extend(onehotStore)
entered_li.append(day_of_the_week)
entered_li.append(promo)
entered_li.extend(onehotState_holiday)
entered_li.append(school_holiday)
entered_li.extend(onehotStore_type)
entered_li.extend(onehotAssortment)
entered_li.append(comp_dist)
entered_li.append(promo2)
entered_li.append(month)
# ========== End of Part 2.3 ==========
# make prediction
prediction = model.predict(np.array(entered_li).reshape(1, -1))
label = str(np.squeeze(prediction.round(2)))
return render_template('index.html', label=label)
def make_prediction():
if request.method == 'POST':
# ========== Part 2.3 ==========
# YOUR CODE START HERE
#******NOTE_******** I was unableto one-hot encode the 'Store' feature in my ipython notebook model so
#I did not use the onehotCategorical function on the 'store' user input_
#NOTE_: I extracted user inputs and one-hot encoded in the same line
# get request values
# one-hot encode categorical variables
store = request.form.get("store")
store_type = onehotCategorical(int(request.form.get("store_type")),
4).astype(str)
assortment = onehotCategorical(int(request.form.get("assortment")),
3).astype(str)
state_holliday = onehotCategorical(
int(request.form.get("state_holliday")), 4).astype(str)
promo2 = request.form.get("promo2")
promo = request.form.get("promo")
day_of_the_week = request.form.get("day_of_the_week")
month = request.form.get("month")
school_holliday = request.form.get("school_holliday")
# manually specify competition distance
comp_dist = '5458.1'
# build 1 observation for prediction
entered_li = np.hstack([
store, store_type, assortment, state_holliday, comp_dist, promo2,
promo, day_of_the_week, month, school_holliday
]).tolist()
# ========== End of Part 2.3 ==========
# make prediction
prediction = model.predict(np.array(entered_li).reshape(1, -1))
label = "$" + str(np.squeeze(prediction.round(2)))
return render_template('index.html', label=label)
def make_prediction():
model = joblib.load('model.pkl')
data = joblib.load('traindatapkl.pkl')
global day
global month
global store
global predictedvalue
if request.method == 'POST':
entered_li = []
month = request.form['Month']
date_entry = month
year, month, day = map(int, date_entry.split('-'))
promo = int(request.form['Promo'])
stateH = int(request.form['StateH'])
schoolH = int(request.form['SchoolH'])
assortment = int(request.form['Assortment'])
storeType = int(request.form['StoreType'])
store = int(request.form['store'])
# one-hot encode categorical variables
stateH_encode = onehotCategorical(stateH, 4)
assortment_encode = onehotCategorical(assortment, 3)
storeType_encode = onehotCategorical(storeType, 4)
store_encode = onehotCategorical(store, 1115, store=1)
comp_dist = 5458.1
entered_li.extend(store_encode)
entered_li.extend(storeType_encode)
entered_li.extend(assortment_encode)
entered_li.extend(stateH_encode)
entered_li.extend([comp_dist])
#entered_li.extend([promo2])
entered_li.extend([promo])
entered_li.extend([day])
entered_li.extend([month])
entered_li.extend([schoolH])
data = [[
store, 1270, promo, schoolH, storeType, assortment, stateH, 6, day,
month, year, 50, 132, 0, 0
]]
df = pd.DataFrame(data,
columns=[
'Store', 'CompetitionDistance', 'Promo',
'SchoolHoliday', 'StoreType', 'Assortment',
'StateHoliday', 'DayOfWeek', 'Month', 'Day',
'Year', 'WeekOfYear', 'CompetitionOpen',
'PromoOpen', 'IsPromoMonth'
])
#data = [[1,1270,1,0,3,1,0,6,9,15,2019,37,132,0,0]]
#df = pd.DataFrame(data,columns=['Store','CompetitionDistance','Promo','SchoolHoliday','StoreType','Assortment','StateHoliday','DayOfWeek','Month','Day','Year','WeekOfYear','CompetitionOpen','PromoOpen','IsPromoMonth'])
prediction = model.predict(xgb.DMatrix(df))
prediction = np.expm1(prediction)
#prediction = model.predict(entered_li.values.reshape(1, -1))
predictedvalue = str(np.squeeze(prediction.round(2)))
predvalueint = np.squeeze(prediction.round(2))
label = "$" + predictedvalue
########################################################
meansales, meansales16 = meansales_of_past_years(store, month)
percentincint = ((predvalueint - meansales) / (predvalueint)) * 100
rawvalue = percentincint.item()
percentinc = str(np.squeeze(percentincint.round(2))) + "%"
if rawvalue > 0:
positive = 1
else:
positive = 0
return render_template('index.html',
label=label,
label1=percentinc,
label2=positive)
def add_user(username):
form = AddUserForm()
#session_user = User.query.filter_by(username=session['username']).first()
session_user = User.query.filter(User.username == username).first()
if request.method == 'GET':
return render_template('add_user.html',
form=form,
title='Add User Information',
session_username=session_user.username)
else:
#session_user.first_name= 'first_name'
#session_user.last_name= 'last_name'
#db.session.commit()
#return redirect('/index')
#session_user = User.query.filter_by(username=session['username']).first()
if form.validate_on_submit():
first_name = request.form['first_name']
last_name = request.form['last_name']
state = request.form['state']
zipcode = request.form['zipcode']
age = request.form['age']
gender = request.form['gender']
drug_type = request.form['drug_type']
use_duration = request.form['use_duration']
med_insurer = request.form['med_insurer']
year_signed_up = request.form['year_signed_up']
years_from_first_litigation = request.form[
'years_from_first_litigation']
copay_or_coinsurance = request.form['copay_or_coinsurance']
personal_spending_per_year = request.form[
'personal_spending_per_year']
income = request.form['income']
if copay_or_coinsurance == 'copay':
total_spending = float(personal_spending_per_year) / .11
else:
total_spending = float(personal_spending_per_year) / .15
#return total_spending
session_user.first_name = first_name
session_user.last_name = last_name
session_user.state = state
session_user.zipcode = zipcode
session_user.age = age
session_user.gender = gender
session_user.drug_type = drug_type
session_user.use_duration = use_duration
session_user.med_insurer = med_insurer
session_user.year_signed_up = year_signed_up
session_user.years_from_first_litigation = years_from_first_litigation
session_user.copay_or_coinsurance = copay_or_coinsurance
session_user.personal_spending_per_year = personal_spending_per_year
session_user.total_spending = total_spending
session_user.income = income
if gender == "M":
gender_1hot = np.array([0, 1])
else:
gender_1hot = np.array([1, 0])
state_1hot = onehotState(state)
years_int = int(years_from_first_litigation)
years_1hot = onehotCategorical(years_int, 20)
print(state_1hot)
#test_pred = np.array([0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 36, 56361, 305.5437079])
#test_pred = np.hstack([gender_1hot, [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], years_1hot, [age, income, total_spending]])
test_pred = np.hstack([
gender_1hot, state_1hot, years_1hot,
[age, income, total_spending]
])
#prediction = model.predict(test_pred.reshape(1, -1))*473.5694972
prediction = model.predict(test_pred.reshape(1,
-1)) * total_spending
session_user.est_settle = float(np.squeeze(prediction.round(2)))
#settle = "$" + str(np.squeeze(prediction.round(2)))
db.session.commit()
#return redirect('/index')
return redirect(url_for('profile', username=session_user.username))
def make_prediction():
if request.method == 'POST':
entered_li = []
# YOUR CODE FOR PART 2.2
# get request values
quarter = int(request.form['Quarter'])
minute = int(request.form['Minute'])
second = int(request.form['Seconds'])
score = int(request.form['Score'])
fav = int(request.form['Favorite'])
home = int(request.form['Home'])
ps = int(request.form['PointSpread'])
half = float(request.form['Half'])
down = int(request.form['Down'])
distance = int(request.form['Distance']) / 10
field = abs(int(request.form['Field']) - 100) / 100
time_left = (3600 - (quarter * 900) + (minute * 60) + second) / 3600
if fav == 1:
posteam_spread = -1 * (ps + half)
else:
posteam_spread = (ps + half)
# one-hot encode categorical variables
down_encode = onehotCategorical(down, 4)
# engineer 1 observation for prediction
# YOUR CODE START HERE
entered_li.extend(down_encode)
entered_li.extend([distance])
entered_li.extend([field])
entered_li.extend([score])
entered_li.extend([home])
entered_li.extend([time_left])
entered_li.extend([posteam_spread])
#pkl_file = open('lr.pkl', 'rb')
#model = pickle.load(pkl_file)
model = joblib.load('lr.pkl')
prediction_o = model.predict_proba(
np.array(entered_li).reshape(1, -1))[0][1] * 100
prediction_d = 100 - prediction_o
#prediction = model.predict(entered_li.values.reshape(1, -1))
if (time_left == 0) & (score > 0):
label_o = '100.00'
label_d = '00.00'
elif (time_left == 0) & (score < 0):
label_o = '00.00'
label_d = '100.00'
elif (0 < time_left < (42 / 3600)) & (score > 0):
label_o = '99.99'
label_d = '0.01'
elif (0 < time_left < (61 / 3600)) & (score > 0):
label_o = '99.49'
label_d = '0.51'
elif (0 < time_left < (42 / 3600)) & (score < -8):
label_o = '0.01'
label_d = '99.99'
elif (0 < time_left < (61 / 3600)) & (score < -8):
label_o = '0.51'
label_d = '99.49'
else:
label_o = str(np.squeeze(prediction_o.round(2)))
label_d = str(np.squeeze(prediction_d.round(2)))
#label_o = str(prediction_o * 100)[:5]
#label_d = str(prediction_d * 100)[:5]
return render_template('index.html', labelo=label_o, labeld=label_d)