shyamsankeerth/Machine_Learning

1import os2import numpy as np3import matplotlib.pyplot as plt4import gradio as gr5import pandas as pd6import tarfile7import urllib.request8 9 10DOWNLOAD_ROOT = "https://raw.githubusercontent.com/ageron/handson-ml2/master/"11HOUSING_PATH = os.path.join("datasets", "housing")12HOUSING_URL = DOWNLOAD_ROOT + "datasets/housing/housing.tgz" 13 14def fetch_housing_data(housing_url=HOUSING_URL, housing_path=HOUSING_PATH):15    if not os.path.isdir(housing_path):16        os.makedirs(housing_path)17    tgz_path = os.path.join(housing_path, "housing.tgz")18    urllib.request.urlretrieve(housing_url, tgz_path)19    housing_tgz = tarfile.open(tgz_path)20    housing_tgz.extractall(path=housing_path)21    housing_tgz.close()22 23 24 25def load_housing_data(housing_path=HOUSING_PATH):26    csv_path = os.path.join(housing_path, "housing.csv")27    return pd.read_csv(csv_path)28 29 30#1. Download the data31 32fetch_housing_data()33 34housing_pd = load_housing_data()35housing_pd.head()36 37## tentatively drop categorical feature38housing = housing_pd.drop('ocean_proximity', axis=1)39housing40 41 42 43#2. Prepare the Data for Machine Learning Algorithms44## 1. split data to get train and test set45from sklearn.model_selection import train_test_split46train_set, test_set = train_test_split(housing, test_size=0.2, random_state=10)47 48## 2. clean the missing values49train_set_clean = train_set.dropna(subset=["total_bedrooms"])50train_set_clean51 52## 2. derive training features and training labels 53train_labels = train_set_clean["median_house_value"].copy() # get labels for output label Y54train_features = train_set_clean.drop("median_house_value", axis=1) # drop labels to get features X for training set55 56 57## 4. scale the numeric features in training set58from sklearn.preprocessing import MinMaxScaler59scaler = MinMaxScaler() ## define the transformer60scaler.fit(train_features) ## call .fit() method to calculate the min and max value for each column in dataset61 62train_features_normalized = scaler.transform(train_features)63train_features_normalized64 65#3. Training ML model on the Training Set66 67from sklearn.linear_model import LinearRegression ## import the LinearRegression Function68lin_reg = LinearRegression() ## Initialize the class69lin_reg.fit(train_features_normalized, train_labels) # feed the training data X, and label Y for supervised learning70 71 72 73### visualize the data74def save_fig(fig_id, tight_layout=True, fig_extension="png", resolution=300):75    path = os.path.join(IMAGES_PATH, fig_id + "." + fig_extension)76    print("Saving figure", fig_id, ' to ',path)77    if tight_layout:78        plt.tight_layout()79    plt.savefig(path, format=fig_extension, dpi=resolution)80    81PROJECT_ROOT_DIR='./'82IMAGES_PATH = os.path.join(PROJECT_ROOT_DIR, "images")83os.makedirs(IMAGES_PATH, exist_ok=True)84 85images_path = os.path.join(PROJECT_ROOT_DIR, "images", "end_to_end_project")86os.makedirs(images_path, exist_ok=True)87DOWNLOAD_ROOT = "https://raw.githubusercontent.com/ageron/handson-ml2/master/"88filename = "california.png"89print("Downloading", filename)90url = DOWNLOAD_ROOT + "images/end_to_end_project/" + filename91urllib.request.urlretrieve(url, os.path.join(images_path, filename))92 93 94### written by Jie95def draw_map_customize(longitude,latitude, fig_id='test',fig_extension='png' ):96    import matplotlib.image as mpimg97    california_img=mpimg.imread(os.path.join(images_path, filename))98    ax = housing.plot(kind="scatter", x="longitude", y="latitude", figsize=(10,7),99                      s=housing['population']/100, label="Population",100                      c="median_house_value", cmap="jet",101                      colorbar=False, alpha=0.4)102    plt.imshow(california_img, extent=[-124.55, -113.80, 32.45, 42.05], alpha=0.5,103               cmap=plt.get_cmap("jet"))104    plt.ylabel("Latitude", fontsize=18)105    plt.xlabel("Longitude", fontsize=18)106 107    plt.xticks(fontsize=18, rotation=0)108    plt.yticks(fontsize=18, rotation=0)109 110 111    plt.plot(longitude,latitude, "ro", alpha=0.7, marker=r'$\clubsuit$', markersize=30)112 113    plt.annotate("Your location is here", xy=(longitude,latitude), xytext=(longitude+1,latitude+1), fontsize=20,114            arrowprops=dict(arrowstyle="->"))115    116    117    prices = housing["median_house_value"]118    tick_values = np.linspace(prices.min(), prices.max(), 11)119    cbar = plt.colorbar(ticks=tick_values/prices.max())120    cbar.ax.set_yticklabels(["$%dk"%(round(v/1000)) for v in tick_values], fontsize=14)121    cbar.set_label('Median House Value', fontsize=16)122    123    plt.legend(fontsize=16)124    save_fig(fig_id)125    #plt.show()126    127    path = os.path.join(IMAGES_PATH, fig_id + "." + fig_extension)128    return path129 130 131 132def get_sample_data(num_data):133    sample_data = []134    for i in range(num_data):135      samp = housing.sample(1)136      longitude = float(samp['longitude'].values[0])137      latitude = float(samp['latitude'].values[0])138      housing_median_age = float(samp['housing_median_age'].values[0])139      total_rooms = float(samp['total_rooms'].values[0])140      total_bedrooms = float(samp['total_bedrooms'].values[0])141      population = float(samp['population'].values[0])142      households = float(samp['households'].values[0])143      median_income = float(samp['median_income'].values[0])144 145      sample_data.append([longitude,latitude,housing_median_age,total_rooms,total_bedrooms,population,households,median_income]) 146    return sample_data147 148 149def predict_price(longitude,latitude,housing_median_age,total_rooms,total_bedrooms,population,households,median_income):150    #import pickle151    #loaded_model = pickle.load(open('KNN_classifier.pickle', 'rb'))152    153    #print(loaded_model)154    # initialize data of lists.155    data = {'longitude':float(longitude),156            'latitude':float(latitude),157            'housing_median_age':float(housing_median_age),158            'total_rooms':float(total_rooms),159            'total_bedrooms':float(total_bedrooms),160            'population':float(population),161            'households':float(households),162            'median_income':float(median_income),163    }164    test_features = pd.DataFrame(columns=['longitude', 'latitude', 'housing_median_age', 'total_rooms',165                      'total_bedrooms', 'population', 'households', 'median_income'])166    # Create DataFrame167    test_features = test_features.append(data,ignore_index=True)168    test_features = test_features.dropna(subset=["total_bedrooms"])169 170    ## 3. scale the numeric features in test set. 171    ## important note: do not apply fit function on the test set, using same scalar from training set172    test_features_normalized = scaler.transform(test_features)173    test_features_normalized174 175    pred = lin_reg.predict(test_features_normalized)[0]176    177    map_file = draw_map_customize(longitude,latitude, fig_id='test',fig_extension='png' )178 179    return pred,map_file180 181 182### configure inputs/outputs183 184 185set_longitude = gr.inputs.Slider(-124.350000, -114.310000, step=0.5, default=-120, label = 'Longitude')186set_latitude = gr.inputs.Slider(32, 41, step=0.5, default=33, label = 'Latitude')187set_housing_median_age = gr.inputs.Slider(1, 52, step=1, default=10, label = 'Housing_median_age (Year)')188set_total_rooms = gr.inputs.Slider(1, 40000, step=5, default=10000, label = 'Total_rooms')189set_total_bedrooms = gr.inputs.Slider(1, 6445, step=5, default=5000, label = 'Total_bedrooms')190set_population = gr.inputs.Slider(3, 35682, step=5, default=10, label = 'Population')191set_households = gr.inputs.Slider(1, 6082, step=5, default=10, label = 'Households')192set_median_income = gr.inputs.Slider(0, 15, step=0.5, default=10, label = 'Median_income')193 194 195 196set_label = gr.outputs.Textbox(label="Predicted Housing Prices")197 198# define output as the single class text199set_out_images = gr.outputs.Image(label="Visualize your location",type="numpy")200 201 202### configure gradio, detailed can be found at https://www.gradio.app/docs/#i_slider203interface = gr.Interface(fn=predict_price, 204                         inputs=[set_longitude, set_latitude,set_housing_median_age,set_total_rooms,set_total_bedrooms,set_population,set_households,set_median_income], 205                         outputs=[set_label,set_out_images],206                         examples_per_page = 2,207                         examples = get_sample_data(10), 208                         title="CSCI4750/5750 Demo 3: Web Application for Housing Price Prediction", 209                         description= "Click examples below for a quick demo",210                         theme = 'huggingface',211                         layout = 'vertical'212                         )213interface.launch(debug=True)