Spaces:
Build error
Build error
File size: 8,560 Bytes
c84fa5d 515a7e5 c84fa5d |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 |
# -*- coding: utf-8 -*-
"""Capstone Gradio App Embedding.ipynb
Automatically generated by Colaboratory.
Original file is located at
https://colab.research.google.com/drive/1zsT_lHGVHzG29XSb4tMF3UdA6glyWnRx
"""
# from google.colab import drive
# drive.mount('/content/drive')
#!pip install gradio
#!pip install category_encoders
"""### **DATA PREP**"""
import pandas as pd
import numpy as np
import gradio as gr
from sklearn.model_selection import train_test_split
from sklearn.model_selection import train_test_split, cross_val_score
from sklearn.metrics import accuracy_score, confusion_matrix, recall_score, precision_recall_curve, f1_score
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import ExtraTreesRegressor
from sklearn.preprocessing import LabelEncoder
import gdown
#train = pd.read_csv('https://drive.google.com/file/d/1U1ic5GE42_cxs8VvuOx7EDaHqtA0oEFC/view?usp=sharing.csv')
#https://drive.google.com/file/d/1U1ic5GE42_cxs8VvuOx7EDaHqtA0oEFC/view?usp=sharing
file_id = '1U1ic5GE42_cxs8VvuOx7EDaHqtA0oEFC'
# Define the URL to download the file
url = f'https://drive.google.com/uc?id={file_id}'
# Download the file and name it 'train.csv'
output_file = 'train.csv'
gdown.download(url, output_file, quiet=False)
# Read the CSV file with Pandas
train = pd.read_csv(output_file)
# use lambda function to remove \t make our model more robst
train = train.applymap(lambda x: x.replace("\t" , '' ) if isinstance (x , str) else x)
# " " , " "
train = train.applymap(lambda x: x.replace(" " , ' ' ) if isinstance (x , str) else x)
# drop what we don't need
train.drop(columns=['MRG', 'user_id', 'ZONE1', 'ZONE2', 'TOP_PACK'], inplace=True)
train["REGION"].fillna(method='ffill', inplace=True)
train["TENURE"].fillna(method='ffill', inplace=True)
train["MONTANT"].fillna(train["MONTANT"].median(), inplace=True)
train["FREQUENCE_RECH"].fillna(0, inplace=True)
train["REVENUE"].fillna(train["REVENUE"].median(), inplace=True)
train["ARPU_SEGMENT"].fillna(0, inplace=True)
train["FREQUENCE"].fillna(0, inplace=True)
train["DATA_VOLUME"].fillna(0, inplace=True)
train["ON_NET"].fillna(0, inplace=True)
train["ORANGE"].fillna(0, inplace=True)
train["TIGO"].fillna(0, inplace=True)
train["REGULARITY"].fillna(train["REGULARITY"].mean(), inplace=True)
train["FREQ_TOP_PACK"].fillna(train["FREQ_TOP_PACK"].mean(), inplace=True)
train['TENURE'] = train['TENURE'].str.replace('D 3-6 month', '1', regex=True)
train['TENURE'] = train['TENURE'].str.replace('E 6-9 month', '2', regex=True)
train['TENURE'] = train['TENURE'].str.replace('F 9-12 month', '3', regex=True)
train['TENURE'] = train['TENURE'].str.replace('J 21-24 month', '4', regex=True)
train['TENURE'] = train['TENURE'].str.replace('G 12-15 month', '5', regex=True)
train['TENURE'] = train['TENURE'].str.replace('H 15-18 month', '6', regex=True)
train['TENURE'] = train['TENURE'].str.replace('I 18-21 month', '7', regex=True)
train['TENURE'] = train['TENURE'].str.replace('K > 24 month', '8', regex=True)
# train['TENURE'].value_counts()
# Define a dictionary to map values
region_mapping = {
'DAKAR': '1',
'THIES': '2',
'SAINT-LOUIS': '3',
'LOUGA': '4',
'KAOLACK': '5',
'DIOURBEL': '6',
'TAMBACOUNDA': '7',
'KAFFRINE': '8',
'KOLDA': '9',
'FATICK': '10',
'ZIGUINCHOR': '11',
'SEDHIOU': '12',
'KEDOUGOU': '13',
'MATAM' : '14'
}
# Use the replace method to map values
train['REGION'] = train['REGION'].replace(region_mapping)
# Look at the new value_counts
# print(train['REGION'].value_counts())
"""## **FITTING AND TRAINING**"""
"""Select target and features"""
y = train['CHURN']
x = train.drop(columns='CHURN', axis=1)
X_train, X_test, y_train, y_test = train_test_split(x,y,test_size = 0.5,random_state=45 )# , stratify=y)
#Further split X_train and y_train into train and validation sets
X_train,X_val,y_train,y_val = train_test_split(X_train,y_train,test_size = 0.3, random_state=1 )#, stratify=y)
"""### SCALE NUMERICAL COLUMNS"""
num_cols = ['MONTANT', 'FREQUENCE_RECH', 'REVENUE', 'ARPU_SEGMENT', 'FREQUENCE',
'DATA_VOLUME', 'ON_NET', 'ORANGE', 'TIGO',
'REGULARITY', 'FREQ_TOP_PACK']
scaler = StandardScaler()
X_train[num_cols] = scaler.fit_transform(X_train[num_cols])
X_val[num_cols] = scaler.fit_transform(X_val[num_cols])
# Create an instance
model = ExtraTreesRegressor(
n_estimators=100, # Number of trees in the forest
max_depth=10, # Maximum depth of the tree
random_state=42 # Random seed for reproducibility
)
# Train the model
MODEL = model.fit(X_train, y_train)
"""## **Check if our model is working**"""
y_pred = MODEL.predict(X_test)
def classifier_1(result):
if result > 0.9:
return "Customer will churn"
else:
return "Customer will not churn"
def predict(REGION,TENURE , MONTANT , FREQUENCE_RECH, REVENUE , ARPU_SEGMENT ,FREQUENCE , DATA_VOLUME , ON_NET, ORANGE , TIGO, REGULARITY ,FREQ_TOP_PACK):
input_array = np.array([[REGION,TENURE , MONTANT , FREQUENCE_RECH, REVENUE , ARPU_SEGMENT ,FREQUENCE , DATA_VOLUME , ON_NET, ORANGE , TIGO, REGULARITY ,FREQ_TOP_PACK]])
pred = MODEL.predict(input_array)
output = classifier_1 (pred[0])
if output == "Customer will churn":
return [(0, output)]
else :
return [(1, output)]
#tenure = tenure_dropdown
REGION = gr.inputs.Slider(minimum=1, maximum=13, label='Location of each client')
TENURE = gr.inputs.Slider(minimum=1, maximum=8, label="Duration in network")
MONTANT = gr.inputs.Slider(minimum=22, maximum=470000, label="Top up amount")
FREQUENCE_RECH = gr.inputs.Slider(minimum=1, maximum=131, label="income frequency")
REVENUE = gr.inputs.Slider(minimum=1, maximum=532177, label="ARPU_SEGMENT")
ARPU_SEGMENT = gr.inputs.Slider(minimum=1, maximum= 177392, label="FREQUENCE")
FREQUENCE = gr.inputs.Slider(minimum=1, maximum=91, label="DATA_VOLUME")
DATA_VOLUME =gr.inputs.Slider(minimum=0, maximum=1702309, label="ON_NET")
ON_NET = gr.inputs.Slider(minimum=0, maximum=36687, label="ORANGE")
ORANGE = gr.inputs.Slider(minimum=0, maximum= 6721, label="TIGO")
TIGO = gr.inputs.Slider(minimum=0, maximum=4174, label="ZONE1")
REGULARITY = gr.inputs.Slider(minimum=1, maximum=62, label="ZONE2")
FREQ_TOP_PACK = gr.inputs.Slider(minimum=1, maximum= 592, label="REGULARITY")
op = gr.outputs.HighlightedText(color_map={"Customer will churn":"pink", "Customer will not churn":"yellow"})
gr.Interface(predict , inputs = [REGION,TENURE, MONTANT , FREQUENCE_RECH, REVENUE , ARPU_SEGMENT ,FREQUENCE , DATA_VOLUME , ON_NET, ORANGE ,TIGO, REGULARITY ,FREQ_TOP_PACK], outputs=op,
live = True).launch(debug=True)
# Input sliders
# REGION = gr.inputs.Slider(minimum=1, maximum=13, label='Location of each client')
# TENURE = gr.inputs.Slider(minimum=1, maximum=8, label="Duration in network")
# MONTANT = gr.inputs.Slider(minimum=22, maximum=470000, label="Top-up amount")
# FREQUENCE_RECH = gr.inputs.Slider(minimum=1, maximum=131, label="Income frequency")
# REVENUE = gr.inputs.Slider(minimum=1, maximum=532177, label="ARPU_SEGMENT")
# ARPU_SEGMENT = gr.inputs.Slider(minimum=1, maximum=177392, label="FREQUENCE")
# FREQUENCE = gr.inputs.Slider(minimum=1, maximum=91, label="DATA_VOLUME")
# DATA_VOLUME = gr.inputs.Slider(minimum=0, maximum=1702309, label="ON_NET")
# ON_NET = gr.inputs.Slider(minimum=0, maximum=36687, label="ORANGE")
# ORANGE = gr.inputs.Slider(minimum=0, maximum=6721, label="TIGO")
# TIGO = gr.inputs.Slider(minimum=0, maximum=4174, label="ZONE1")
# REGULARITY = gr.inputs.Slider(minimum=1, maximum=62, label="ZONE2")
# FREQ_TOP_PACK = gr.inputs.Slider(minimum=1, maximum=592, label="REGULARITY")
# # Output configuration
# op = gr.outputs.HighlightedText(color_map={"Customer will churn": "pink", "Customer will not churn": "yellow"})
# # Create and launch the interface
# gr.Interface(predict, inputs=[REGION, TENURE, MONTANT, FREQUENCE_RECH, REVENUE, ARPU_SEGMENT, FREQUENCE,
# DATA_VOLUME, ON_NET, ORANGE, TIGO, REGULARITY, FREQ_TOP_PACK], outputs=op,
# live=False).launch(debug=False)
# # Map numerical values to labels
# tenure_labels = {
# 0: "3-6 months",
# 1: "6-9 months",
# 2: "9-12 months",
# 3: "12-15 months",
# 4: "15-18 months",
# 5: "18-21 months",
# 6: "21-24 months",
# 7: "> 24 months"
# }
# # Reverse the mapping for predictions
# tenure_values = {v: k for k, v in tenure_labels.items()}
# # Create a dropdown menu with labels
# tenure_dropdown = gr.inputs.Dropdown(list(tenure_labels.values()), label="TENURE")
|