def get_cluster_coords(location, price, ptype, psize):
# get clusters,listings etc
clusters = get_data('clusters')
#listings = get_data('listings')
coordinates = cluster_coordinates(location, price, ptype, psize, clusters)
return coordinates
def get_cluster_ranking(location, price, ptype, psize, listing_id):
# get clusters,listings etc
clusters = get_data('clusters')
#listings = get_data('listings')
cl_rank, cl_average, cl_scores = user_ranking(location, price, ptype, psize, listing_id, clusters)
return cl_rank, cl_average, cl_scores
def get_wordcloud(cluster_id):
# the dataframe wordcounts_df contains the most common 2 words associations
# and associated counts for the clusters in label_list as well as
# across all clusters (if the cluster label for listing_id is not in this
# list, the wordcloud will be based on the word counts calculated across
# all clusters)
label_list = ['L:Westminster_P:very_cheap_S:room',\
'L:Tower Hamlets_P:very_cheap_S:room',\
'L:Tower Hamlets_P:cheap_S:room',\
'L:Camden_P:very_cheap_S:room',\
'L:Lambeth_P:cheap_S:room',\
'L:Camden_P:cheap_S:room',\
'L:Southwark_P:cheap_S:room',\
'L:Hammersmith and Fulham_P:very_cheap_S:room',\
'L:Southwark_P:very_cheap_S:room',\
'L:Lambeth_P:very_cheap_S:room',\
'L:Westminster_P:cheap_S:small',\
'L:Hackney_P:cheap_S:room',\
'L:Islington_P:very_cheap_S:room',\
'L:Kensington and Chelsea_P:very_cheap_S:room',\
'L:Westminster_P:cheap_S:room',\
'L:Islington_P:cheap_S:room',\
'L:Westminster_P:very_expensive_S:large',\
'L:Hackney_P:very_cheap_S:room',\
'L:Westminster_P:expensive_S:large',\
'L:Westminster_P:average_S:small',\
'L:Wandsworth_P:cheap_S:room']
wordcounts_df = get_data('wordcount')
# If cluster_id not in above list, set to 'All'. The wordcloud will be based
# on the word associations calculated across all clusters
if cluster_id not in label_list:
cluster_id = 'All'
# Extract word_counts from wordcl_df for cluster_id
word_counts = wordcounts_df.loc[wordcounts_df['cluster'] == cluster_id]
# Turn into right format for wordcloud and create wordcloud
word_counts = pd.Series(word_counts['count'].to_list(),
index=word_counts['quotes'].to_list())
wordcloud = WordCloud(
background_color="white",
collocation_threshold=5).generate_from_frequencies(word_counts)
return wordcloud
def listing_to_cluster(listing_id):
clusters = get_data('clusters')
cluster_id = clusters[clusters['listing_id']==listing_id].iloc[0]['cluster']
return cluster_id
def __init__(self):
self.listings = get_data()
self.clusters = get_data('clusters')
self.model = Model().load_model()
self.build_cluster_info()