Adding final chapters files

Chapter9/dash-example-no-user-inputs.py

@@ -0,0 +1,50 @@
+# -*- coding: utf-8 -*-
+"""
+@author: Alvaro Fuentes
+Chapter 9. Hands-On Predictive Analytics with Python
+Building a basic static app
+"""
+## imports
+import dash
+import dash_core_components as dcc
+import dash_html_components as html
+import plotly.graph_objs as go
+import pandas as pd
+import os
+
+## Importing the dataset
+DATA_DIR = '../data'
+FILE_NAME = 'diamonds.csv'
+data_path = os.path.join(DATA_DIR, FILE_NAME)
+diamonds = pd.read_csv(data_path)
+
+## Creating the app
+app = dash.Dash(__name__)
+
+# Creating a Plotly figure
+trace = go.Histogram(
+        x = diamonds['price']
+        )
+
+layout = go.Layout(
+        title = 'Diamond Prices',
+        xaxis = dict(title='Price'),
+        yaxis = dict(title='Count')
+        )
+
+figure = go.Figure(
+        data = [trace],
+        layout = layout
+        )
+
+app.layout = html.Div([
+        html.H1('My first Dash App'),
+        html.H2('Histogram of diamond prices'),
+        html.P('This is some normal text, we can use it to describe something about the application.'),          
+        dcc.Graph(id='my-histogram', figure=figure)
+        ])
+
+
+if __name__ == '__main__':
+    app.run_server(debug=True) 
+

Chapter9/dash-example-user-inputs.py

@@ -0,0 +1,92 @@
+# -*- coding: utf-8 -*-
+"""
+@author: Alvaro Fuentes
+Chapter 9. Hands-On Predictive Analytics with Python
+Building a basic interactive app
+"""
+## imports
+import dash
+import dash_core_components as dcc
+import dash_html_components as html
+from dash.dependencies import Input, Output
+import plotly.graph_objs as go
+import pandas as pd
+import os
+
+## Importing the dataset
+DATA_DIR = '../data'
+FILE_NAME = 'diamonds.csv'
+data_path = os.path.join(DATA_DIR, FILE_NAME)
+diamonds = pd.read_csv(data_path)
+diamonds = diamonds.sample(n=2000)
+
+
+app = dash.Dash(__name__)
+
+app.css.append_css({
+    'external_url': 'https://codepen.io/chriddyp/pen/bWLwgP.css'
+})
+
+numerical_features = ['price','carat','depth','table','x','y','z']
+options_dropdown = [{'label':x.upper(), 'value':x} for x in numerical_features]
+
+dd_x_var = dcc.Dropdown(
+        id='x-var',
+        options = options_dropdown,
+        value = 'carat'
+        )
+
+div_x_var =  html.Div(
+        children=[html.H4('Variable for x axis: '), dd_x_var],
+        className="six columns"
+        )
+
+
+dd_y_var = dcc.Dropdown(
+        id='y-var',
+        options = options_dropdown,
+        value = 'price'
+        ) 
+
+div_y_var =  html.Div(
+        children=[html.H4('Variable for y axis: '), dd_y_var],
+        className="six columns"
+        )
+
+app.layout = html.Div(children=[
+        html.H1('Adding interactive controls'),
+        html.H2('Interactive scatter plot example'),
+        html.Div(
+                children=[div_x_var, div_y_var],
+                className="row"
+                ),  
+        dcc.Graph(id='scatter')
+        ])
+
+
+@app.callback(
+        Output(component_id='scatter', component_property='figure'),
+        [Input(component_id='x-var', component_property='value'), Input(component_id='y-var', component_property='value')])
+def scatter_plot(x_col, y_col):
+    trace = go.Scatter(
+            x = diamonds[x_col],
+            y = diamonds[y_col],
+            mode = 'markers'
+            )
+
+    layout = go.Layout(
+            title = 'Scatter plot',
+            xaxis = dict(title = x_col.upper()),
+            yaxis = dict(title = y_col.upper())
+            )
+
+    output_plot = go.Figure(
+            data = [trace],
+            layout = layout
+            )
+
+    return output_plot
+
+
+if __name__ == '__main__':
+    app.run_server(debug=True)

Chapter9/diamonds-model-training.py

@@ -0,0 +1,74 @@
+# -*- coding: utf-8 -*-
+"""
+@author: Alvaro Fuentes
+Chapter 9. Hands-On Predictive Analytics with Python
+Producing the predictive model's objects
+"""
+## Imports
+import numpy as np
+import pandas as pd
+import os
+from keras.models import Sequential
+from keras.layers import Dense
+from sklearn.externals import joblib
+
+## Loading the dataset
+DATA_DIR = '../data'
+FILE_NAME = 'diamonds.csv'
+data_path = os.path.join(DATA_DIR, FILE_NAME)
+diamonds = pd.read_csv(data_path)
+
+
+## Preparing the dataset
+diamonds = diamonds.loc[(diamonds['x']>0) | (diamonds['y']>0)]
+diamonds.loc[11182, 'x'] = diamonds['x'].median()
+diamonds.loc[11182, 'z'] = diamonds['z'].median()
+diamonds = diamonds.loc[~((diamonds['y'] > 30) | (diamonds['z'] > 30))]
+diamonds = pd.concat([diamonds, pd.get_dummies(diamonds['cut'], prefix='cut', drop_first=True)], axis=1)
+diamonds = pd.concat([diamonds, pd.get_dummies(diamonds['color'], prefix='color', drop_first=True)], axis=1)
+diamonds = pd.concat([diamonds, pd.get_dummies(diamonds['clarity'], prefix='clarity', drop_first=True)], axis=1)
+
+## Dimensionality reduction
+from sklearn.decomposition import PCA
+pca = PCA(n_components=1, random_state=123)
+diamonds['dim_index'] = pca.fit_transform(diamonds[['x','y','z']])
+diamonds.drop(['x','y','z'], axis=1, inplace=True)
+
+## Creating X and y
+X = diamonds.drop(['cut','color','clarity','price'], axis=1)
+y = np.log(diamonds['price'])
+
+## Standarization: centering and scaling
+numerical_features = ['carat', 'depth', 'table', 'dim_index']
+from sklearn.preprocessing import StandardScaler
+scaler = StandardScaler()
+X.loc[:, numerical_features] = scaler.fit_transform(X[numerical_features])
+
+## Building the neural network
+n_input = X.shape[1]
+n_hidden1 = 32
+n_hidden2 = 16
+n_hidden3 = 8
+
+nn_reg = Sequential()
+nn_reg.add(Dense(units=n_hidden1, activation='relu', input_shape=(n_input,)))
+nn_reg.add(Dense(units=n_hidden2, activation='relu'))
+nn_reg.add(Dense(units=n_hidden3, activation='relu'))
+# output layer
+nn_reg.add(Dense(units=1, activation=None))
+
+## Training the neural network
+batch_size = 32
+n_epochs = 40
+nn_reg.compile(loss='mean_absolute_error', optimizer='adam')
+nn_reg.fit(X, y, epochs=n_epochs, batch_size=batch_size)
+
+## Serializing:
+# PCA
+joblib.dump(pca, './Model/pca.joblib') 
+
+# Scaler
+joblib.dump(scaler, './Model/scaler.joblib')
+
+# Trained model
+nn_reg.save("./Model/diamond-prices-model.h5")