app.py

######################
# Import libraries
######################

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import streamlit as st
import seaborn as sns
from PIL import Image
import base64

st.set_option('deprecation.showPyplotGlobalUse', False)

icon = Image.open("icon.ico")
st.set_page_config(
     page_title="Kickstarter Web App",
     layout="wide",
     page_icon=icon,
     initial_sidebar_state="expanded",
 )

######################
# Page Title
######################

# Image

image = Image.open("logo_app.jpg")

st.image(image, use_column_width=True)

# Title

st.title("Kickstarter Web App")

st.write("This app helps to improve your crowdfunding project!")

# About section

expander_bar = st.expander("About")
expander_bar.markdown("""
* **Python libraries:** numpy, pandas, streamlit, matplotlib, seaborn, base64, PIL
* **Data source:** [Kaggle.com](https://www.kaggle.com/kemical/kickstarter-projects)
* **App Developers:** M.Majdak, D.Migoń, P.Rakowska [github.com](https://github.com/infoshareacademy/paradoma)
""")


# Divide page to 3 columns (col1 = sidebar, col2 and col3 = page contents)
col1 = st.sidebar
col2, col3 = st.columns((3,1))

######################
# Import data
######################

@st.cache
def load_data():
    df = pd.read_csv("ks_projects_201801.csv")
    # Drop NaN
    df = df.dropna()
    # Convert data to datetime format
    df['deadline_new'] = pd.to_datetime(df['deadline']).dt.date
    df['launched_new'] = pd.to_datetime(df['launched']).dt.date
    # Calculating the duration of projects, cleaning data - conversion to the int type 
    df['duration'] = df['deadline_new'] - df['launched_new']
    df['duration'] = df['duration'].map(lambda x: int(str(x).split(" ")[0]))
    # Extract two project states: successful = 1, others = 0
    df["state_new"] = df["state"].map(lambda x: 1 if x == "successful" else 0 )
    # Removal of unnecessary columns
    to_drop = ["currency",
            "deadline",
            "goal",
             "launched",
            "pledged",
            "usd pledged"]
    df.drop(to_drop, inplace=True, axis=1)
    # Joining Luxembourg, Belgium and the Netherlands data as the Benelux
    df["country"] = df["country"].replace(["LU", "BE"],"NL")
    # Removal of countries with less than 1000 projects
    df = df[(df["country"] != "SG") 
            & (df["country"] != "JP") 
            & (df["country"] != "AT") 
            & (df["country"] != "HK")
            & (df["country"] != "NO")
            & (df["country"] != "CH")
            & (df["country"] != "IE")]
    # Extracting the year and month from the data 
    df['year_launched'] = pd.DatetimeIndex(df['launched_new']).year
    df['month_launched'] = pd.DatetimeIndex(df['launched_new']).month
    # Removal of records from 1970
    df = df[df["year_launched"] != 1970]

    return df

df = load_data()

######################
# Input - country selection
######################

col2.header("Choose Your Country and See the Stats")
col2.text("")

country_list = ["Australia", "Benelux", "Canada", "Denmark", "France", "Germany", "Great Britain", "Italy", "Mexico", "New Zealand", "Spain", "Sweden", "United States"]
country_dict = {
    "Australia": "AU",
    "Benelux": "NL",
    "Canada": "CA",
    "Denmark": "DK",
    "France": "FR",
    "Germany": "DE",
    "Great Britain": "GB",
    "Italy": "IT",
    "Mexico": "MX",
    "New Zealand": "NZ",
    "Spain": "ES",
    "Sweden": "SE",
    "United States": "US"
}

selected_country = col2.selectbox('Country', country_list)
selected_country_conv = country_dict[selected_country]

######################
# Output - after country selection
######################

col2.text("")
if col2.button("Country Stats"):

    col2.text("")
    col2.subheader("Details about Kickstarter crowdfunding in " + selected_country)
    col2.text("")

    # Number of projects each year
    country_filter = df[df["country"] == selected_country_conv]
    data_all = country_filter.groupby(by="year_launched").size().reset_index(name="sum")
    success_filter = country_filter[country_filter["state_new"] == 1]
    data_success = success_filter.groupby(by="year_launched").size().reset_index(name="sum_success")
    data = pd.merge(data_all, data_success, on="year_launched", how="left")
    plt.subplots(figsize=(10, 5))
    plt.xlabel("Year", fontdict={"fontsize": 15})
    plt.ylabel("Qty of projects", fontdict={"fontsize": 15})
    plt.xticks(fontsize=13)
    plt.yticks(fontsize=13)
    plt.title("Number of projects each year", fontdict={"fontname": "monospace", "fontsize": 18})
    plt.xlim(2008.5,2018.5)
    plt.bar(data["year_launched"], data["sum"], color="palegreen", width=0.5, fill=True, linestyle=":", edgecolor="#000000", label="Total")
    plt.bar(data["year_launched"], data["sum_success"], color="limegreen", width=0.5, fill=True, linestyle=":", edgecolor="#000000", label="Succeed")
    plt.legend()
    col2.pyplot()

    # Average duration of the project
    avg_dur = round(country_filter["duration"].mean(), 1)
    col2.text("")
    col2.markdown(f"""<span style='color:darkgreen; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid green;'> Average duration of the project in {selected_country}: <b>{avg_dur}</b> days </span>""", unsafe_allow_html=True)

    # Percent of successful projects
    success_filter = country_filter[country_filter["state"]=="successful"]
    perc_success = round(len(success_filter)/len(country_filter)*100, 1)
    col2.text("")
    col2.markdown(f"""<span style='color:darkgreen; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid green;'> Percentage of successful projects in this country: <b>{perc_success}%</b>.</span>""", unsafe_allow_html=True)
    
    # The best project (the highest amount of money)
    pro_bes = country_filter[["ID", "name", "deadline_new", "usd_pledged_real", "backers", "main_category"]]
    project_best = country_filter["usd_pledged_real"].idxmax()
    project_best_dict = pro_bes.loc[project_best].to_dict()
    col2.text("")
    col2.markdown(f"""<span style='color:limegreen; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid palegreen;'> The most money was raised for the <b>"{project_best_dict["name"]}"</b> project.</span>""", unsafe_allow_html=True)
    
    col2.text("")
    image_2 = Image.open("png/" + selected_country_conv +".png")
    col2.image(image_2, width=800)
    
    col2.text("")
    col2.markdown(f"""<span style='color:limegreen; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid palegreen;'> The collection finished successfully on <b>{project_best_dict["deadline_new"]}</b>.</span>""", unsafe_allow_html=True) 
    col2.text("")
    col2.markdown(f"""<span style='color:limegreen; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid palegreen;'> Project creators raised $ <b>{project_best_dict["usd_pledged_real"]}</b>.</span>""", unsafe_allow_html=True)
    col2.text("")
    col2.markdown(f"""<span style='color:limegreen; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid palegreen;'> The project had <b>{project_best_dict["backers"]}</b> founders and was submitted in the <b>{project_best_dict["main_category"]}</b> category.</span>""", unsafe_allow_html=True)

    # The best project (the largest number of backers)
    top_backers = pro_bes["backers"].idxmax()
    top_backers_dict = pro_bes.loc[top_backers].to_dict()
    
    col2.text("")
    col2.markdown(f"""<span style='color:goldenrod; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid gold;'> The project with the largest number of backers was <b>"{top_backers_dict["name"]}"</b>. It had <b>{top_backers_dict['backers']}</b> founders.</span>""", unsafe_allow_html=True)
    
    col2.text("")
    image_3 = Image.open("png/" + selected_country_conv +"b.png")
    col2.image(image_3, width=800)

    # Sum backers in selected country
    sum_backers = country_filter["backers"].sum()
    
    col2.text("")
    col2.markdown(f"""<span style='color:darkgreen; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid green;'> The total number of people supporting crowdfunding projects in this country is <b>{sum_backers}</b>.</span>""", unsafe_allow_html=True)
    col2.text("")

    # The most popular category
    category_filter = country_filter.groupby("main_category")
    m_cat_list = category_filter["main_category"].count().sort_values(ascending=False).head(1).to_dict()
    
    for cat, num in m_cat_list.items():
        col2.markdown(f"""<span style='color:darkgreen; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid green;'> The most popular category in {selected_country} is <b>{cat}</b> (<b>{num}</b> projects).</span>""", unsafe_allow_html=True)
        
    col2.text("")
    
    # The most successful main category
    category_filter_succes = country_filter[country_filter["state"]=="successful"]
    category_filter_succes = category_filter_succes.groupby("main_category")
    m_cat_list2 = category_filter_succes["main_category"].count().sort_values(ascending=False).head(1).to_dict()
    
    for cat2, num2 in m_cat_list2.items():
        col2.markdown(f"""<span style='color:darkgreen; font-family:Georgia; font-size:1.5em; display: block; text-align:center; border: 2px solid green;'> The category with the highest number of successful projects is <b>{cat2}</b>. There were <b>{num2}</b> completed  projects.</span>""", unsafe_allow_html=True)

    col2.text("")

######################
# Input - params selection
######################

col2.markdown("""<span style='color:darkred; font-family:"Comic Sans MS"; font-size:2em; display: block; text-align:center;'> <b>🠈 To find out what are the chances for your project to succeed, choose params in the sidebar.</b></span>""", unsafe_allow_html=True)
col2.text("")

# Sidebar
col1.header("Input Options")

# Main category selection
sorted_unique_m_category = sorted(df["main_category"].unique())
selected_m_category = col1.multiselect("Main category", sorted_unique_m_category, sorted_unique_m_category[:1])

# Category selection (dependent on main category)
main_cat_select = df[df["main_category"].isin(selected_m_category)]

sorted_unique_category = sorted(main_cat_select["category"].unique())
selected_category = col1.multiselect("Category", sorted_unique_category, sorted_unique_category)

# Project duration selection
selected_duration = col1.slider("How many days is the fundraising going to last?", min_value=1, max_value=100)
    
# Project amount selection
selected_amount = col1.number_input("How much money do you want to collect? [$]", min_value=1, step=10)

######################
# Output - after params selection
######################

# Filtering data
cf = df[df["country"] == selected_country_conv]

df_selected_projects = df[(df["country"] == selected_country_conv) & 
                        (df["main_category"].isin(selected_m_category)) & 
                        (df["category"].isin(selected_category))]

interval_list = [(0, 10), (10, 20), (20, 30), (30, 40), (40, 50), (50, 60), (60, 70), (70, 80), (80, 90), (90, 100)]
amount_list = list(cf["usd_goal_real"].quantile([0.2, 0.4, 0.6, 0.8]))

for start, stop in interval_list:
    if selected_duration > start and selected_duration <= stop:
        if selected_amount < amount_list[0]:
            df_selected_projects = df[(df["country"] == selected_country_conv) & 
                                    (df["main_category"].isin(selected_m_category)) & 
                                    (df["category"].isin(selected_category)) &
                                    (cf["duration"] > start) & (cf["duration"] <= stop) &
                                    (cf["usd_goal_real"] < amount_list[0])]
        elif selected_amount >= amount_list[0] and selected_amount < amount_list[1]:
            df_selected_projects = df[(df["country"] == selected_country_conv) & 
                                    (df["main_category"].isin(selected_m_category)) & 
                                    (df["category"].isin(selected_category)) &
                                    (cf["duration"] > start) & (cf["duration"] <= stop) &
                                    (cf["usd_goal_real"] >= amount_list[0]) & (cf["usd_goal_real"] < amount_list[1])]
        elif selected_amount >= amount_list[1] and selected_amount < amount_list[2]:
            df_selected_projects = df[(df["country"] == selected_country_conv) & 
                                    (df["main_category"].isin(selected_m_category)) & 
                                    (df["category"].isin(selected_category)) &
                                    (cf["duration"] > start) & (cf["duration"] <= stop) &
                                    (cf["usd_goal_real"] >= amount_list[1]) & (cf["usd_goal_real"] < amount_list[2])]
        elif selected_amount >= amount_list[2] and selected_amount < amount_list[3]:
            df_selected_projects = df[(df["country"] == selected_country_conv) & 
                                    (df["main_category"].isin(selected_m_category)) & 
                                    (df["category"].isin(selected_category)) &
                                    (cf["duration"] > start) & (cf["duration"] <= stop) &
                                    (cf["usd_goal_real"] >= amount_list[2]) & (cf["usd_goal_real"] < amount_list[3])] 
        elif selected_amount >= amount_list[3]:
            df_selected_projects = df[(df["country"] == selected_country_conv) & 
                                    (df["main_category"].isin(selected_m_category)) & 
                                    (df["category"].isin(selected_category)) &
                                    (cf["duration"] > start) & (cf["duration"] <= stop) &
                                    (cf["usd_goal_real"] >= amount_list[3])]

col2.header("Display Projects Info with Selected Params")
col2.success("Data Quantity: " + str(df_selected_projects.shape[0]) + " projects (" + str(df_selected_projects[df_selected_projects["state_new"] == 1].shape[0]) + " successful).")
col2.dataframe(df_selected_projects[["name", "category", 
                                    "main_category", "state", 
                                    "backers", "country", 
                                    "usd_pledged_real", 
                                    "usd_goal_real", 
                                    "launched_new", "duration"]].rename(
                                        columns= {
                                            "main_category": "main category",
                                            "usd_pledged_real" : "collected [$]",
                                            "usd_goal_real": "goal [$]",
                                            "launched_new": "launched",
                                            "duration": "duration [days]"
                                        }))

# Download selected projects data
# https://discuss.streamlit.io/t/how-to-download-file-in-streamlit/1806

def filedownload(df):
    csv = df.to_csv(index=False)
    b64 = base64.b64encode(csv.encode()).decode()  # strings <-> bytes conversions
    href = f'<a href="data:file/csv;base64,{b64}" download="projectinfo.csv">Download CSV File</a>'
    return href

col2.markdown(filedownload(df_selected_projects), unsafe_allow_html=True)

######################
# Output - probability calc
######################

col2.text("")
col2.header("Probability of getting funds")

def mcat_prob():
    # Founded projects in selected main category
    mcat_s = cf[(cf["main_category"].isin(selected_m_category)) & (cf["state_new"] == 1)]
    # No fundation in selected main category
    mcat_no_s = cf[(cf["main_category"].isin(selected_m_category)) & (cf["state_new"] == 0)]
    # The probability that the project will be funded if it is submitted to a certain main category
    prob_main_category = round(100 * mcat_s["state_new"].count() / (mcat_s["state_new"].count() + mcat_no_s["state_new"].count()), 1)
    return prob_main_category

def cat_prob():      
    # Founded projects in selected category
    cat_s = cf[(cf["category"].isin(selected_category)) & (cf["state_new"] == 1)]
    # No fundation in selected category
    cat_no_s = cf[(cf["category"].isin(selected_category)) & (cf["state_new"] == 0)]
    # The probability that the project will be funded if it is submitted to a certain category
    prob_category = round(100 * cat_s["state_new"].count() / (cat_s["state_new"].count() + cat_no_s["state_new"].count()), 1)
    return prob_category    

# The probability that the project will be funded if it lasts a certain amount of time
prob_duration = 100 * (cf[cf["state_new"] == 1].count()/cf["state_new"].count())
# Checks in which range is the value specified by the user
interval_list = [(0, 10), (10, 20), (20, 30), (30, 40), (40, 50), (50, 60), (60, 70), (70, 80), (80, 90), (90, 100)]
for start, stop in interval_list:
    if selected_duration > start and selected_duration <= stop:
         # Founded projects in selected duration
        dur_s = cf[ (cf["duration"] > start) & (cf["duration"] <= stop) & (cf["state_new"] == 1) ]
        # No fundation in selected duration
        dur_no_s = cf[ (cf["duration"] > start) & (cf["duration"] <= stop) & (cf["state_new"] == 0) ]
        # The probability that the project will be funded if it lasts for a certain amount of time
        prob_duration = round(100 * dur_s["state_new"].count() / (dur_s["state_new"].count() + dur_no_s["state_new"].count()), 1)

# The probability that the project will be funded depending on the amount of money needed
prob_amount = 100 * (cf[cf["state_new"] == 1].count()/cf["state_new"].count())

# Checks in which range is the value specified by the user
amount_list = list(cf["usd_goal_real"].quantile([0.2, 0.4, 0.6, 0.8]))
if selected_amount < amount_list[0]:   
    amount_s = cf[ (cf["usd_goal_real"] < amount_list[0]) & (cf["state_new"] == 1) ]
    amount_no_s = cf[ (cf["usd_goal_real"] < amount_list[0]) & (cf["state_new"] == 0) ]
    prob_amount = round(100 * amount_s["state_new"].count() / (amount_s["state_new"].count() + amount_no_s["state_new"].count()), 1)     
elif selected_amount >= amount_list[0] and selected_amount < amount_list[1]: 
    amount_s = cf[ (cf["usd_goal_real"] >= amount_list[0]) & (cf["usd_goal_real"] < amount_list[1]) & (cf["state_new"] == 1)]
    amount_no_s = cf[ (cf["usd_goal_real"] >= amount_list[0]) & (cf["usd_goal_real"] < amount_list[1]) & (cf["state_new"] == 0) ]
    prob_amount = round(100 * amount_s["state_new"].count() / (amount_s["state_new"].count() + amount_no_s["state_new"].count()), 1)
elif selected_amount >= amount_list[1] and selected_amount < amount_list[2]:
    amount_s = cf[ (cf["usd_goal_real"] >= amount_list[1]) & (cf["usd_goal_real"] < amount_list[2]) & (cf["state_new"] == 1)]
    amount_no_s = cf[ (cf["usd_goal_real"] >= amount_list[1]) & (cf["usd_goal_real"] < amount_list[2]) & (cf["state_new"] == 0) ]
    prob_amount = round(100 * amount_s["state_new"].count() / (amount_s["state_new"].count() + amount_no_s["state_new"].count()), 1)  
elif selected_amount >= amount_list[2] and selected_amount < amount_list[3]:
    amount_s = cf[ (cf["usd_goal_real"] >= amount_list[2]) & (cf["usd_goal_real"] < amount_list[3]) & (cf["state_new"] == 1)]
    amount_no_s = cf[ (cf["usd_goal_real"] >= amount_list[2]) & (cf["usd_goal_real"] < amount_list[3]) & (cf["state_new"] == 0) ]
    prob_amount = round(100 * amount_s["state_new"].count() / (amount_s["state_new"].count() + amount_no_s["state_new"].count()), 1)
elif selected_amount >= amount_list[3]:       
    amount_s = cf[(cf["usd_goal_real"] >= amount_list[3]) & (cf["state_new"] == 1)]
    amount_no_s = cf[(cf["usd_goal_real"] >= amount_list[3]) & (cf["state_new"] == 0) ]
    prob_amount = round(100 * amount_s["state_new"].count() / (amount_s["state_new"].count() + amount_no_s["state_new"].count()), 1)

# The probalility that the project will be funded depending on all factors
p_abcd = np.sum((df["country"] == selected_country_conv) & 
                (df["main_category"].isin(selected_m_category)) & 
                (df["category"].isin(selected_category))) / cf["ID"].count()

p_abcde = np.sum((df["country"] == selected_country_conv) & 
                (df["main_category"].isin(selected_m_category)) & 
                (df["category"].isin(selected_category)) &
                (df["state_new"] == 1)) / cf["ID"].count()

for start, stop in interval_list:
    if selected_duration > start and selected_duration <= stop:
        if selected_amount < amount_list[0]:
            p_abcd = np.sum((df["country"] == selected_country_conv) & 
                            (df["main_category"].isin(selected_m_category)) & 
                            (df["category"].isin(selected_category)) &
                            (cf["duration"] > start) & (cf["duration"] <= stop) &
                            (cf["usd_goal_real"] < amount_list[0])) / cf["ID"].count()
            p_abcde = np.sum((df["country"] == selected_country_conv) & 
                             (df["main_category"].isin(selected_m_category)) & 
                             (df["category"].isin(selected_category)) &
                             (df["state_new"] == 1) &
                             (cf["duration"] > start) & (cf["duration"] <= stop) &
                             (cf["usd_goal_real"] < amount_list[0])) / cf["ID"].count()
        elif selected_amount >= amount_list[0] and selected_amount < amount_list[1]:
            p_abcd = np.sum((df["country"] == selected_country_conv) & 
                            (df["main_category"].isin(selected_m_category)) & 
                            (df["category"].isin(selected_category)) &
                            (cf["duration"] > start) & (cf["duration"] <= stop) &
                            (cf["usd_goal_real"] >= amount_list[0]) & (cf["usd_goal_real"] < amount_list[1])) / cf["ID"].count()
            p_abcde = np.sum((df["country"] == selected_country_conv) & 
                             (df["main_category"].isin(selected_m_category)) & 
                             (df["category"].isin(selected_category)) &
                             (df["state_new"] == 1) &
                             (cf["duration"] > start) & (cf["duration"] <= stop) &
                             (cf["usd_goal_real"] >= amount_list[0]) & (cf["usd_goal_real"] < amount_list[1])) / cf["ID"].count()
        elif selected_amount >= amount_list[1] and selected_amount < amount_list[2]:
            p_abcd = np.sum((df["country"] == selected_country_conv) & 
                            (df["main_category"].isin(selected_m_category)) & 
                            (df["category"].isin(selected_category)) &
                            (cf["duration"] > start) & (cf["duration"] <= stop) &
                            (cf["usd_goal_real"] >= amount_list[1]) & (cf["usd_goal_real"] < amount_list[2])) / cf["ID"].count()
            p_abcde = np.sum((df["country"] == selected_country_conv) & 
                             (df["main_category"].isin(selected_m_category)) & 
                             (df["category"].isin(selected_category)) &
                             (df["state_new"] == 1) &
                             (cf["duration"] > start) & (cf["duration"] <= stop) &
                             (cf["usd_goal_real"] >= amount_list[1]) & (cf["usd_goal_real"] < amount_list[2])) / cf["ID"].count()
        elif selected_amount >= amount_list[2] and selected_amount < amount_list[3]:
            p_abcd = np.sum((df["country"] == selected_country_conv) & 
                            (df["main_category"].isin(selected_m_category)) & 
                            (df["category"].isin(selected_category)) &
                            (cf["duration"] > start) & (cf["duration"] <= stop) &
                            (cf["usd_goal_real"] >= amount_list[2]) & (cf["usd_goal_real"] < amount_list[3])) / cf["ID"].count()
            p_abcde = np.sum((df["country"] == selected_country_conv) & 
                             (df["main_category"].isin(selected_m_category)) & 
                             (df["category"].isin(selected_category)) &
                             (df["state_new"] == 1) &
                             (cf["duration"] > start) & (cf["duration"] <= stop) &
                             (cf["usd_goal_real"] >= amount_list[2]) & (cf["usd_goal_real"] < amount_list[3])) / cf["ID"].count()
        elif selected_amount >= amount_list[3]:
            p_abcd = np.sum((df["country"] == selected_country_conv) & 
                            (df["main_category"].isin(selected_m_category)) & 
                            (df["category"].isin(selected_category)) &
                            (cf["duration"] > start) & (cf["duration"] <= stop) &
                            (cf["usd_goal_real"] >= amount_list[3])) / cf["ID"].count()
            p_abcde = np.sum((df["country"] == selected_country_conv) & 
                             (df["main_category"].isin(selected_m_category)) & 
                             (df["category"].isin(selected_category)) &
                             (df["state_new"] == 1) &
                             (cf["duration"] > start) & (cf["duration"] <= stop) &
                             (cf["usd_goal_real"] >= amount_list[3])) / cf["ID"].count()

if p_abcde == 0 or p_abcd == 0:
    answear = "### <span style='color:darkred; font-size:1.5em'>The overal probability cannot be determined!!!</span>"
else:
    p = round((p_abcde / p_abcd) * 100, 1)
    answear = (f"""### The overall probability of getting funds is <span style='color:limegreen; font-family:"Comic Sans MS"; font-size:4em; display: block; text-align:center;'><b> {p}% </b></span>.""")
    
# Response to the user

col2.text("")
if col2.button("Calculate"):
    col2.text("")
    col2.markdown(f"""
    #### Probability of funding a project in selected main categories: 
    <span style='color:limegreen; font-family:"Comic Sans MS"; font-size:2.5em; display: block; text-align:center;'><b> {mcat_prob()}% </b></span>.
    """, unsafe_allow_html=True)
    col2.text("")
    col2.markdown(f"""
    #### Probability of funding a project in selected categories: 
    <span style='color:limegreen; font-family:"Comic Sans MS"; font-size:2.5em; display: block; text-align:center;'><b> {cat_prob()}% </b></span>.""", unsafe_allow_html=True)
    col2.text("")
    col2.markdown(f"""
    #### Probability of funding for project duration {selected_duration} days: 
    <span style='color:limegreen; font-family:"Comic Sans MS"; font-size:2.5em; display: block; text-align:center;'><b> {prob_duration}% </b></span>.""", unsafe_allow_html=True)
    col2.text("")
    col2.markdown(f"""
    #### Probability of financing a project with an estimated cost of $ {selected_amount}: 
    <span style='color:limegreen; font-family:"Comic Sans MS"; font-size:2.5em; display: block; text-align:center;'><b> {prob_amount}% </b></span>.""", unsafe_allow_html=True)
    col2.text("")
    col2.markdown(answear, unsafe_allow_html=True)

######################
# Output - improving project
######################

col2.text("")
col2.header("To improve your project, click below")
col2.text("")
if col2.button("I want to succeed"):
    
    # Main category
    col2.text("")
    col2.markdown("""<span style='color:darkgreen; font-family:Georgia; font-size:2em; display: block; text-align:center;'> <b>MAIN CATEGORY</b></span>.""", unsafe_allow_html=True)
    col2.text("")
    
    for one_mcat in selected_m_category:
        category_filter = cf[cf["main_category"] == one_mcat]
        success_filter = category_filter[category_filter['state_new']==1]
        
        if len(success_filter) > 0 and len(category_filter) > 0:
            a = round(len(success_filter)/len(category_filter)*100, 1)
            col2.markdown(f"""### Success rate of your main category <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{one_mcat}</b></span> is <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{a}%</b></span>.""", unsafe_allow_html=True)
        else:
            a = 0
            col2.markdown(f"<span style='color:darkred; font-family:Calibri; font-size:1em;'> <b>It's not possible to calculate success rate of your main category {one_mcat}</b>.</span>""", unsafe_allow_html=True)
        
        list_category= cf["main_category"].unique().tolist()
        list_category_graph_x = [one_mcat]
        list_category_graph_y = [a]
        info_2 = []
        graph= 0
        for i in list_category:
            category_filter_a = cf[cf["main_category"] == i]
            success_filter_a = category_filter_a[category_filter_a["state_new"] == 1]
            if len(success_filter_a) > 0 and len(category_filter_a) > 0:
                b = round(len(success_filter_a)/len(category_filter_a)*100, 1)
                if b > a:
                    graph +=1
                    info_1 = """<span style='color:darkred; font-family:Calibri; font-size:1.5em; display: block; text-align:center;'> <b>Think about other main categories with higher success rate in your country (of course if applicable!):</b></span>"""
                    info_2.append(f"##### * {i} - {b}%")
                    list_category_graph_x.append(i)
                    list_category_graph_y.append(b)
                else: 
                    info_1 = ""#"<span style='color:darkred; font-family:Calibri; font-size:2em; display: block; text-align:center;'> <b>Congratulations! You choose the most successful main category in your country!</b> </span>"""
        col2.markdown(info_1, unsafe_allow_html=True)
        if len(info_2) > 0:
            for info in info_2:
                col2.markdown(info)
        if graph > 0:
            sns.set_theme(style="whitegrid", context="talk")

            plt.figure(figsize=(10, 5))
            plt.ylabel("Success Rate [%]", fontdict={'fontsize': 15})
            plt.xlabel("", fontdict={"fontsize": 15})
            plt.xticks(rotation=90, fontsize=13) 
            plt.yticks(fontsize=13)
            sns.barplot(x=list_category_graph_x, y=list_category_graph_y, palette="crest")
            col2.pyplot()

    # Category
    col2.markdown(""""<span style='color:darkgreen; font-family:Georgia; font-size:2em; display: block; text-align:center;'> <b>CATEGORY</b></span>.""", unsafe_allow_html=True)
    for one_cat in selected_category:
        subcategory_filter = category_filter[category_filter["category"] == one_cat]
        success_filter_2 = subcategory_filter[subcategory_filter["state_new"] == 1]
    
        if len(success_filter_2) > 0 and len(subcategory_filter) > 0:
            c = round(len(success_filter_2)/len(subcategory_filter)*100, 1)
            col2.markdown(f"""### Success rate of your category <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{one_cat}</b></span> is <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{c}%</b></span>.""", unsafe_allow_html=True)
        else:
            c = 0
            col2.markdown(f"""<span style='color:darkred; font-family:Calibri; font-size:1.5em;'> <b>It's not possible to calculate success rate of your category {one_cat}</b>.</span>""", unsafe_allow_html=True)

        list_subcategory = category_filter["category"].unique().tolist()
        list_subcategory_graph_x = [one_cat]
        list_subcategory_graph_y = [c]
        info_4 = []
        graph_2= 0

        for i in list_subcategory:
            subcategory_filter_a = category_filter[category_filter["category"] == i]
            success_filter_2_a = subcategory_filter_a[subcategory_filter_a["state_new"] == 1]
            if len(success_filter_2_a) > 0 and len(subcategory_filter_a) > 0:
                d = round(len(success_filter_2_a)/len(subcategory_filter_a)*100, 1)
                if d > c:
                    graph_2 +=1
                    info_3 = """<span style='color:darkred; font-family:Calibri; font-size:1.5em; display: block; text-align:center;'> <b>Think about other categories with higher success rate in your country (of course if applicable!):</b></span>"""
                    info_4.append(f"##### * {i} - {d}%")
                    list_subcategory_graph_x.append(i)
                    list_subcategory_graph_y.append(d)
                else: 
                    info_3 = ""#"<span style='color:darkred; font-family:Calibri; font-size:2em; display: block; text-align:center;'> <b>Congratulations! You choose the most successful category in your country!</b> </span>"""
        
        col2.markdown(info_3, unsafe_allow_html=True)
        if len(info_4) > 0:
            for info in info_4:
                col2.markdown(info)
        
        if graph_2 > 0:
            sns.set_theme(style="whitegrid", context="talk")

            plt.figure(figsize=(10, 5))
            plt.ylabel("Success Rate [%]", fontdict={'fontsize': 15})
            plt.xlabel("", fontdict={"fontsize": 15})
            plt.xticks(rotation=90, fontsize=13) 
            plt.yticks(fontsize=13)
            sns.barplot(x=list_subcategory_graph_x, y=list_subcategory_graph_y, palette="crest")
            col2.pyplot()

    # Duration
    for start, stop in interval_list:
        if selected_duration > start and selected_duration <= stop:
            time_y = start
            time_x = stop
    
    col2.markdown("""<span style='color:darkgreen; font-family:Georgia; font-size:2em; display: block; text-align:center;'> <b>TIME</b></span>.""", unsafe_allow_html=True)
    time_filter = category_filter[category_filter["duration"].between(time_y, time_x)]
    success_filter_3 = time_filter[time_filter["state_new"] == 1]
    if len(time_filter) > 0 and len(success_filter_3) > 0:
        e = round(len(success_filter_3)/len(time_filter)*100, 1)
        col2.markdown(f"""### Success rate of your anticipated time range of <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{time_y} - {time_x}</b></span> days is <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{e}%</b></span>.""", unsafe_allow_html=True)
        col2.text("")
        col2.text("")
        col2.text("")
    else:
        e = 0
        col2.markdown("""<span style='color:darkred; font-family:Calibri; font-size:1.5em; display: block; text-align:center;'> <b> It's not possible to calculate success rate of selected time range</b>.</span>""", unsafe_allow_html=True)
    
    better_t = 0
    list_time_graph_x =[]
    list_time_graph_y =[]
    for field in interval_list:
        time_filter_T = category_filter[category_filter["duration"].between(field[0], field[1])]
        success_filter_T = time_filter_T[time_filter_T["state_new"] == 1]
        if len(time_filter_T) > 0 and len(success_filter_T) > 0:
            f = round(len(success_filter_T)/len(time_filter_T)*100, 1)
            list_time_graph_x.append("{}-{}".format(field[0], field[1]))
            list_time_graph_y.append(f)
            if e < f:
                better_t += 1
                col2.markdown(f"""#### Projects in your main category with anticipated time range of <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{field[0]} - {field[1]}</b></span> days are characterized by success rate of <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{f}%</b></span>.""", unsafe_allow_html=True)
                
                if field[1] > time_x:
                    col2.markdown("""<span style='color:darkred; font-family:Calibri; font-size:1.5em; display: block; text-align:center;'> <b> Reconsider increasing the anticipated time</b>.</span>""", unsafe_allow_html=True)
                elif field[1] < time_x:
                    col2.markdown("""<span style='color:darkred; font-family:Calibri; font-size:1.5em; display: block; text-align:center;'> <b> Reconsider decreassing the anticipated time</b>.</span>""", unsafe_allow_html=True)
            elif e == f:
                continue
            else:
                col2.markdown(f"#### Projects in your main category with anticipated time range of <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{field[0]} - {field[1]}</b></span> days are characterized by success rate of <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{f}%</b></span>.""", unsafe_allow_html=True)
                col2.markdown("""<span style='color:darkred; font-family:Calibri; font-size:1.5em; display: block; text-align:center;'> <b> The result is worse than ours. Let's not go that way</b>.</span>""", unsafe_allow_html=True)
    col2.markdown("""##### This might be helpful:
* [crowdfundingpr.org](https://www.crowdfundingpr.org/ways-boost-kickstarter-campaign/)
* [kickstarter.com](https://www.kickstarter.com/blog/shortening-the-maximum-project-length/)""")
    if better_t == 0:
        col2.markdown("""<span style='color:darkred; font-family:Calibri; font-size:2em; display: block; text-align:center;'> <b> SEEMS THAT YOU SELECTED THE BEST RANGE!</b></span>""", unsafe_allow_html=True)    
    elif better_t > 0: 
       
        sns.set_theme(style="whitegrid", context="talk")

        plt.figure(figsize=(10, 5))
        plt.ylabel("Success Rate [%]", fontdict={'fontsize': 15})
        plt.xlabel("Days", fontdict={"fontsize": 15})
        plt.xticks(fontsize=13) 
        plt.yticks(fontsize=13)
        sns.barplot(x=list_time_graph_x, y=list_time_graph_y, palette="crest")
        col2.pyplot()

    # Money
    if selected_amount < amount_list[0]:
        money_y = 0
        money_x = round(amount_list[0], 2)
    elif selected_amount >= amount_list[0] and selected_amount < amount_list[1]:
        money_y = round(amount_list[0], 2)
        money_x = round(amount_list[1], 2)
    elif selected_amount >= amount_list[1] and selected_amount < amount_list[2]:
        money_y = round(amount_list[1], 2)
        money_x = round(amount_list[2], 2)
    elif selected_amount >= amount_list[2] and selected_amount < amount_list[3]:
        money_y = round(amount_list[2], 2)
        money_x = round(amount_list[3], 2)  
    elif selected_amount >= amount_list[3]:
        money_y = round(amount_list[3], 2)
        money_x = 25000000 

    col2.markdown(""""<span style='color:darkgreen; font-family:Georgia; font-size:2em; display: block; text-align:center;'> <b>MONEY</b></span>.""", unsafe_allow_html=True)
    money_filter = category_filter[category_filter["usd_goal_real"].between(money_y, money_x)]
    success_filter_3 = money_filter[money_filter["state_new"] == 1]
    if len(money_filter) > 0 and len(success_filter_3) > 0:
        g = round(len(success_filter_3)/len(money_filter)*100, 1)
        col2.markdown(f"### Success rate of your anticipated money range of <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{money_y} - {money_x}</b></span> $ is <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{g}%</b></span>.""", unsafe_allow_html=True)
        col2.text("")
        col2.text("")
        col2.text("")

    money_ranges= [(0,round(amount_list[0], 2)),(round(amount_list[0], 2),round(amount_list[1], 2)),(round(amount_list[1], 2),round(amount_list[2], 2)), (round(amount_list[2], 2),round(amount_list[3], 2)),(round(amount_list[3], 2),25000000) ]
    better_m = 0
    list_money_graph_x = []
    list_money_graph_y =[]
    for money in money_ranges:
        money_filter_P = category_filter[category_filter["usd_goal_real"].between(money[0], money[1], inclusive="left")]
        success_filter_P = money_filter_P[money_filter_P["state_new"] == 1]
        if len(money_filter_P) > 0 and len(success_filter_P) > 0:
            h = round(len(success_filter_P)/len(money_filter_P)*100, 1)
            list_money_graph_x.append("{} - {}".format(round(money[0], 2), round(money[1], 2)))
            list_money_graph_y.append(h)
            if g < h:
                col2.markdown(f"#### Projects in your category with anticipated money range of <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{round(money[0], 2)} - {round(money[1], 2)}</b></span> $ are characterized by success rate of: <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{h}%</b></span>.""", unsafe_allow_html=True)
                better_m += 1
                if money[1] > money_x:
                    col2.markdown("""<span style='color:darkred; font-family:Calibri; font-size:1.5em; display: block; text-align:center;'> <b> Reconsider increasing the anticipated amount of money to be collected</b>.</span>""", unsafe_allow_html=True)
                elif money[1] < money_x:
                    col2.markdown("""<span style='color:darkred; font-family:Calibri; font-size:1.5em; display: block; text-align:center;'> <b> Reconsider decreassing the anticipated amount of money to be collected</b>.</span>""", unsafe_allow_html=True)
            elif g == h:
                continue
            else:
                col2.markdown(f"#### Projects in your category with anticipated money range of <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{round(money[0], 2)} - {round(money[1], 2)}</b></span> $ are characterized by success rate of: <span style='color:darkgreen; font-family:Calibri; font-size:1em;'><b>{h}%</b></span>.""", unsafe_allow_html=True)
                col2.markdown("""<span style='color:darkred; font-family:Calibri; font-size:1.5em; display: block; text-align:center;'> <b> The result is worse than ours. Let's not go that way</b>.</span>""", unsafe_allow_html=True)
    if better_m == 0:
        col2.markdown("""<span style='color:darkred; font-family:Calibri; font-size:2em; display: block; text-align:center;'> <b> SEEMS THAT YOU SELECTED THE BEST RANGE!</b></span>""", unsafe_allow_html=True)
    elif better_m > 0:
        sns.set_theme(style="whitegrid", context="talk")

        plt.figure(figsize=(10, 5))
        plt.ylabel("Success Rate [%]", fontdict={'fontsize': 15})
        plt.xlabel("Anticipated $ Range", fontdict={"fontsize": 15})
        plt.xticks(rotation=90, fontsize=13) 
        plt.yticks(fontsize=13)
        sns.barplot(x=list_money_graph_x, y=list_money_graph_y, palette="crest")
        col2.pyplot()

######################
# THE END
######################