app.py

# import packages 
import streamlit as st
import joblib
import pandas as pd
import numpy as np
from os.path import dirname, join, realpath
import joblib


# add banner image
st.header("Tanzania Tourism Expenditure Prediction")
st.image("images/tanzania-mount-kilimanjaro.jpg")
st.subheader(
    """
A simple machine learning app to predict how much money a tourist will spend when visiting Tanzania.
"""
)

# form to collect Tourist information
my_form = st.form(key="financial_form")


@st.cache
# function to trasform Yes and No options
def func(value):
    if value == 1:
        return "Yes"
    else:
        return "No"


country = my_form.selectbox(
    "select country",
    (
        "SWIZERLAND",
        "UNITED KINGDOM",
        "CHINA",
        "SOUTH AFRICA",
        "UNITED STATES OF AMERICA",
        "NIGERIA",
        "INDIA",
        "BRAZIL",
        "CANADA",
        "MALT",
        "MOZAMBIQUE",
        "RWANDA",
        "AUSTRIA",
        "MYANMAR",
        "GERMANY",
        "KENYA",
        "ALGERIA",
        "IRELAND",
        "DENMARK",
        "SPAIN",
        "FRANCE",
        "ITALY",
        "EGYPT",
        "QATAR",
        "MALAWI",
        "JAPAN",
        "SWEDEN",
        "NETHERLANDS",
        "UAE",
        "UGANDA",
        "AUSTRALIA",
        "YEMEN",
        "NEW ZEALAND",
        "BELGIUM",
        "NORWAY",
        "ZIMBABWE",
        "ZAMBIA",
        "CONGO",
        "BURGARIA",
        "PAKISTAN",
        "GREECE",
        "MAURITIUS",
        "DRC",
        "OMAN",
        "PORTUGAL",
        "KOREA",
        "SWAZILAND",
        "TUNISIA",
        "KUWAIT",
        "DOMINICA",
        "ISRAEL",
        "FINLAND",
        "CZECH REPUBLIC",
        "UKRAIN",
        "ETHIOPIA",
        "BURUNDI",
        "SCOTLAND",
        "RUSSIA",
        "GHANA",
        "NIGER",
        "MALAYSIA",
        "COLOMBIA",
        "LUXEMBOURG",
        "NEPAL",
        "POLAND",
        "SINGAPORE",
        "LITHUANIA",
        "HUNGARY",
        "INDONESIA",
        "TURKEY",
        "TRINIDAD TOBACCO",
        "IRAQ",
        "SLOVENIA",
        "UNITED ARAB EMIRATES",
        "COMORO",
        "SRI LANKA",
        "IRAN",
        "MONTENEGRO",
        "ANGOLA",
        "LEBANON",
        "SLOVAKIA",
        "ROMANIA",
        "MEXICO",
        "LATVIA",
        "CROATIA",
        "CAPE VERDE",
        "SUDAN",
        "COSTARICA",
        "CHILE",
        "NAMIBIA",
        "TAIWAN",
        "SERBIA",
        "LESOTHO",
        "GEORGIA",
        "PHILIPINES",
        "IVORY COAST",
        "MADAGASCAR",
        "DJIBOUT",
        "CYPRUS",
        "ARGENTINA",
        "URUGUAY",
        "MORROCO",
        "THAILAND",
        "BERMUDA",
        "ESTONIA",
        "BOTSWANA",
        "VIETNAM",
        "GUINEA",
        "MACEDONIA",
        "HAITI",
        "LIBERIA",
        "SAUD ARABIA",
        "BOSNIA",
        "BULGARIA",
        "PERU",
        "BANGLADESH",
        "JAMAICA",
        "SOMALI",
    ),
)

age_group = my_form.selectbox("Select you age range", ("1-24", "25-44", "45-64", "65+"))

travel_with = my_form.selectbox(
    "Who do you plan to travel with?",
    ("Friends/Relatives", "Alone", "Spouse", "Children", "Spouse and Children"),
)
purpose = my_form.selectbox(
    "What is the purpose of visiting Tanzania?",
    (
        "Leisure and Holidays",
        "Visiting Friends and Relatives",
        "Business",
        "Meetings and Conference",
        "Volunteering",
        "Scientific and Academic",
        "Other",
    ),
)

total_number = my_form.number_input(
    "How many people are you are coming with in Tanzania?", min_value=1
)

main_activity = my_form.selectbox(
    "What is the main activity you want to do when you are in Tanzania?",
    (
        "Wildlife tourism",
        "Cultural tourism",
        "Mountain climbing",
        "Beach tourism",
        "Conference tourism",
        "Hunting tourism",
        "Bird watching",
        "business",
        "Diving and Sport Fishing",
    ),
)

tour_arrangement = my_form.selectbox(
    "How do your arrange your tour?", ("Independent", "Package Tour")
)

package_transport_int = my_form.selectbox(
    "Does the package tour include International Transportation?",
    (0, 1),
    format_func=func,
)
package_accomodation = my_form.selectbox(
    "Does the package tour include Accomodation service?", (0, 1), format_func=func,
)
package_food = my_form.selectbox(
    "Does the package tour include Food service?", (0, 1), format_func=func
)
package_transport_tz = my_form.selectbox(
    "Does the package tour include Local Transportation when you are in Tanzania?",
    (0, 1),
    format_func=func,
)
package_sightseeing = my_form.selectbox(
    "Does the package tour include Sight Seeing service?", (0, 1), format_func=func
)
package_guided_tour = my_form.selectbox(
    "Does the package tour include Tour guiding service?", (0, 1), format_func=func
)
package_insurance = my_form.selectbox(
    "Does the package tour include Insurance?", (0, 1), format_func=func
)
payment_mode = my_form.selectbox(
    "What is your payment mode for tourism service?",
    ("Cash", "Credit Card", "Other", "Travellers Cheque"),
)
first_trip_tz = my_form.selectbox(
    "Is this your first trip to Tanzania?", (0, 1), format_func=func
)

night_mainland = my_form.number_input(
    "How many days you plan to spend in Tanzania Mainland", min_value=0,
)
night_zanzibar = my_form.number_input(
    "How many days you plan to spend in Zanzibar", min_value=0
)

submit = my_form.form_submit_button(label="make prediction")


# load the model and one-hot-encoder and scaler

with open(
    join(dirname(realpath(__file__)), "model/histgradient-tanzania-tourism-model.pkl"),
    "rb",
) as f:
    model = joblib.load(f)

with open(join(dirname(realpath(__file__)), "preprocessing/scaler.pkl"), "rb") as f:
    scaler = joblib.load(f)


with open(
    join(dirname(realpath(__file__)), "preprocessing/one-hot-encoder.pkl"), "rb"
) as f:
    one_hot_encoder = joblib.load(f)

# result dictionary
result_dic = {
    1: " from Tsh 0 to Tsh 500,000",
    2: "from Tsh 500,001 to Tsh 1,000,000",
    3: "from Tsh 1,000,001 to Tsh 5,000,000",
    4: "from Tsh 5,000,001 to Tsh 10,000,000",
    5: "from Tsh 10,000,001 and above",
}


@st.cache
# function to clean and tranform the input
def preprocessing_data(data, one_hot_enc, scaler):

    # For other variables let's use one-hot-encoder
    multi_categorical_variables = [
        "country",
        "age_group",
        "tour_arrangement",
        "travel_with",
        "purpose",
        "main_activity",
        "payment_mode",
    ]

    multi_categorical_data = data[multi_categorical_variables]

    multi_categorical_data = one_hot_enc.transform(multi_categorical_data)

    data = data.drop(multi_categorical_variables, axis=1)

    data = data.to_numpy()

    final_data = np.concatenate([data, multi_categorical_data], axis=1)

    final_data = scaler.transform(final_data)

    return final_data


if submit:

    # collect inputs
    input = {
        "country": country,
        "age_group": age_group,
        "travel_with": travel_with,
        "total_number": total_number,
        "purpose": purpose,
        "main_activity": main_activity,
        "tour_arrangement": tour_arrangement,
        "package_transport_int": package_accomodation,
        "package_accomodation": package_accomodation,
        "package_food": package_food,
        "package_transport_tz": package_transport_tz,
        "package_sightseeing": package_sightseeing,
        "package_guided_tour": package_guided_tour,
        "package_insurance": package_insurance,
        "night_mainland": night_mainland,
        "night_zanzibar": night_zanzibar,
        "payment_mode": payment_mode,
        "first_trip_tz": first_trip_tz,
    }

    # create a dataframe
    data = pd.DataFrame(input, index=[0])

    # clean and transform input
    transformed_data = preprocessing_data(
        data=data, one_hot_enc=one_hot_encoder, scaler=scaler
    )

    # perform prediction
    prediction = model.predict(transformed_data)
    output = int(prediction[0])

    # Display results of the Tourism prediction
    st.header("Results")
    st.write(" You are expected to spend:{} ".format(result_dic[output]))


url = "https://twitter.com/Davis_McDavid"
st.write("Developed with ❤️ by [Davis David](%s)" % url)