Ricardo ricardomurillo

## In [31]
recommend('Hampton Inn Suites National HarborAlexandria Area')

## In [25]
recommend('Silver Sands Oceanfront Motel')

## In [21]
recommend('AC Hotel Miami Beach')

## In [18]
def recommend(hotel_name):
    # get index of hotel in ratings matrix
    index = ratings_matrix.columns.get_loc(hotel_name)
    # person coefficients of the hotel
    hotel_coeff = pearson_matrix[index]
    # create dataframe and sort by coeff and get top 10
    return pd.DataFrame({'pearson':hotel_coeff, 'hotel': ratings_matrix.columns}).sort_values('pearson', ascending=False).head(1

## In [8]
pearson_matrix = np.corrcoef(ratings_svd)
pearson_matrix.shape

## in [7]
from sklearn.decomposition import TruncatedSVD

SVD = TruncatedSVD(n_components=10, random_state=34)
ratings_svd = SVD.fit_transform(ratings_transpose)
ratings_svd.shape

## TEST
import pandas as pd
import numpy as np
import sklearn

## In [6]
ratings_transpose = ratings_matrix.T

ratings_transpose.head(5)

## In [5]
ratings_matrix = ratings.pivot_table(values='rating', index='user', columns='hotel', fill_value=0)
ratings_matrix.head(5)

## In[4]
df = pd.read_csv('hotel_reviews.csv')
df.head()

ratings = df[['reviews.username','name','reviews.rating']]
ratings = ratings.drop_duplicates()
ratings.columns = ["user","hotel","rating"]

#8866 unique ratings
ratings.head(5)
	def recommend(hotel_name):
	# get index of hotel in ratings matrix
	index = ratings_matrix.columns.get_loc(hotel_name)
	# person coefficients of the hotel
	hotel_coeff = pearson_matrix[index]
	# create dataframe and sort by coeff and get top 10
	return pd.DataFrame({'pearson':hotel_coeff, 'hotel': ratings_matrix.columns}).sort_values('pearson', ascending=False).head(1
	pearson_matrix = np.corrcoef(ratings_svd)
	pearson_matrix.shape
	from sklearn.decomposition import TruncatedSVD

	SVD = TruncatedSVD(n_components=10, random_state=34)
	ratings_svd = SVD.fit_transform(ratings_transpose)
	ratings_svd.shape
	ratings_transpose = ratings_matrix.T

	ratings_transpose.head(5)
	ratings_matrix = ratings.pivot_table(values='rating', index='user', columns='hotel', fill_value=0)
	ratings_matrix.head(5)
	df = pd.read_csv('hotel_reviews.csv')
	df.head()

	ratings = df[['reviews.username','name','reviews.rating']]
	ratings = ratings.drop_duplicates()
	ratings.columns = ["user","hotel","rating"]

	#8866 unique ratings
	ratings.head(5)