#librerias import numpy as np import pandas as pd import seaborn as sns import datetime from sklearn.preprocessing import StandardScaler from sklearn.model_selection import train_test_split from sklearn import datasets, linear_model from sklearn.preprocessing import PolynomialFeatures # extraccion de los datos dataframe = pd.read_csv('avocado-updated-2020.csv') dataframe.head()

#Estadisticas del dataframe dataframe.describe()

print(dataframe.columns) #cabeceras print(dataframe.dtypes) # tipos de datos

#organicacion y acondicionamiento de los datos #Se eliminan algunos datos de los cuales no se tiene informacion clara # se elimina el dato año ya que contiene un formato extraño que no correspone a los datos reales df = dataframe.drop(['total_bags','small_bags','large_bags','xlarge_bags','year'], axis=1) df['date']= pd.to_datetime(df['date'], errors='coerce')#converion de tipo de la columna date object to datetime df['year']= df['date'].dt.year # extaccion del año df['month']= df['date'].dt.month # extaccion del mes df['day']= df['date'].dt.day # extraccion del dia #Se convierten los datos de tipo y ubicacion a datos numericos convirtiendolos en columnas tipo =pd.get_dummies(df['type']) ubicacion=pd.get_dummies(df['geography']) df2 =pd.concat((df,tipo,ubicacion), axis=1) # Se eliminan los datos tipo texto df2 = df2.drop(['type','geography'],axis=1) df2

print(df2.dtypes) # tipos de datos print(df2.columns) # columnas

#Media media =df2['average_price'].mean() #Mediana mediana = df2['average_price'].median() print('media = {} \nmediana = {}'.format(media, mediana))

#Histograma de frecuencia df2['average_price'].plot.hist(bins=20) #bins = intervalos

#distribution plot, histograma sns.displot(df2, x='average_price', hue=('organic')) #hue crea un histograma por cada tipo de aguacate #0 conventional 1 organic

#desviacion estandar desviacion_estandar=df2['average_price'].std() #Rango = valor max - valor min rango = df2['average_price'].max() - df2['average_price'].min() #Quartiles mediana = df2['average_price'].median() Q1 = df2['average_price'].quantile(q=0.25) #25% de los datos Q3 = df2['average_price'].quantile(q=0.75) #75% de los datos min_val = df2['average_price'].min() max_val = df2['average_price'].max() print('minimo ={min_val}\nQuartil(Q1) ={Q1}\nmediana ={mediana}\nQuartil(Q3) ={Q3}\nmaximo ={max_val}\n'.format( min_val=min_val, Q1=Q1, mediana=mediana, Q3=Q3, max_val=max_val))

#limites de deteccion de valores atipicos iqr = Q3 - Q1 minlimit = Q1 - 1.5*iqr maxlimit = Q3 + 1.5*iqr print(iqr,minlimit, maxlimit)

sns.histplot(df2['average_price'])

# diagrama de caja sns.boxplot(df2['average_price'])

scaler = StandardScaler() scaled = scaler.fit_transform( df2[['total_volume', '4046', '4225', '4770', 'year', 'month', 'day']].values ) covariance_matrix = np.cov(scaled.T) covariance_matrix eigen_values, eigen_vectors = np.linalg.eig(covariance_matrix) variance_explained = [] for i in eigen_values: variance_explained.append((i/sum(eigen_values))*100) print(variance_explained)

from sklearn.decomposition import PCA pca = PCA(n_components=7) pca.fit(scaled) pca.explained_variance_ratio_

reduced_scaled = pca.transform(scaled) df2['pca_1'] = reduced_scaled[:,0] df2['pca_2'] = reduced_scaled[:,1] df2['pca_3'] = reduced_scaled[:,2] df2['pca_4'] = reduced_scaled[:,3] df2['pca_5'] = reduced_scaled[:,4] df2['pca_6'] = reduced_scaled[:,5] df2['pca_7'] = reduced_scaled[:,6] df2.head() df2.columns

y = df2['average_price'] x = df2[[ 'conventional', 'organic', 'Albany', 'Atlanta', 'Baltimore/Washington', 'Boise', 'Boston', 'Buffalo/Rochester', 'California', 'Charlotte', 'Chicago', 'Cincinnati/Dayton', 'Columbus', 'Dallas/Ft. Worth', 'Denver', 'Detroit', 'Grand Rapids', 'Great Lakes', 'Harrisburg/Scranton', 'Hartford/Springfield', 'Houston', 'Indianapolis', 'Jacksonville', 'Las Vegas', 'Los Angeles', 'Louisville', 'Miami/Ft. Lauderdale', 'Midsouth', 'Nashville', 'New Orleans/Mobile', 'New York', 'Northeast', 'Northern New England', 'Orlando', 'Philadelphia', 'Phoenix/Tucson', 'Pittsburgh', 'Plains', 'Portland', 'Raleigh/Greensboro', 'Richmond/Norfolk', 'Roanoke', 'Sacramento', 'San Diego', 'San Francisco', 'Seattle', 'South Carolina', 'South Central', 'Southeast', 'Spokane', 'St. Louis', 'Syracuse', 'Tampa', 'Total U.S.', 'West', 'West Tex/New Mexico', 'pca_1', 'pca_2', 'pca_3', 'pca_4', 'pca_5','pca_6']] pf = PolynomialFeatures(degree = 2) X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.2) X_train_poli = pf.fit_transform(X_train) X_test_poli = pf.fit_transform(X_test) lr_multiple = linear_model.LinearRegression() lr_multiple.fit(X_train_poli, y_train) Y_pred_multiple = lr_multiple.predict(X_test_poli) print('DATOS DEL MODELO REGRESIÓN LINEAL MULTIPLE') print() print('Valor de las pendientes o coeficientes "a":') print(lr_multiple.coef_) print('Valor de la intersección o coeficiente "b":') print(lr_multiple.intercept_) print('Precisión del modelo:') print(lr_multiple.score(X_train_poli, y_train))