Spotify Countries Proximity

import hashlib import io import numpy as np import pandas as pd import requests countries = [ "ar", "cl", "us", "es", "gb" ] def get_download_url(country): return f"https://spotifycharts.com/regional/{country}/daily/latest/download" def get_country_df(country): country_url = get_download_url(country) r = requests.get(country_url) csv_binary = r.content.decode("utf-8") return pd.read_csv(io.StringIO(csv_binary), header=1, error_bad_lines=False).head(10) def get_all_countries_tracks_dict(): countries_dict = {} for country in countries: try: df = pd.read_csv(f'regional-{country}-weekly-latest.csv', header=1) tracks = list(df["Track Name"]) if tracks: countries_dict[country] = tracks except Exception as E: print(E) print(f"{country} encoding failure") return countries_dict def get_all_tracks_as_set(countries_dict): unique_tracks = set() for country in list(countries_dict.keys()): unique_tracks.update(countries_dict[country]) return unique_tracks def LD(s, t): if s == "": return len(t) if t == "": return len(s) if s[-8:] == t[-8:]: cost = 0 else: cost = 1 res = min([LD(s[:-8], t)+1, LD(s, t[:-8])+1, LD(s[:-8], t[:-8]) + cost]) return res def encode_set(unique_tracks): hash_dict = {} for track in unique_tracks: hash_dict[track] = str(abs(hash(track))) return hash_dict def get_all_countries_tracks_dict_hashed(countries_dict, all_tracks_encoded_dict): countries_dict_hashed = {} for country in countries_dict: countries_dict_hashed[country] = [all_tracks_encoded_dict[track] for track in countries_dict[country]] return countries_dict_hashed def LD(s, t): if s == "": return len(t) if t == "": return len(s) if s[-8:] == t[-8:]: cost = 0 else: cost = 1 res = min([LD(s[:-8], t)+1, LD(s, t[:-8])+1, LD(s[:-8], t[:-8]) + cost]) return res def minimumEditDistance(first, second): #Creating numpy ndarray( initialized with 0 of dimension of size of both strings len_first = int(len(first)/19) len_second = int(len(second)/19) matrix = np.zeros((len_first+1, len_second+1), dtype=np.int) # Cross relation loop through each character of each string with each other and # fill the respective index of matrxi (row,column) for i in range(len_first+1): for j in range(len_second+1): #First doing the boundary value analysis, if first or second string is empty so directly adding insertion cost if i == 0: matrix[i][j] = j #Second case elif j == 0: matrix[i][j] = i else: x = i - 1 + 19 * (i - 1) y = i - 1 + 19 * (i - 1 + 1) matrix[i][j] = min(matrix[i][j-1] + 1, matrix[i-1][j] + 1, matrix[i-1][j-1] + 1 if first[x:y] != second[x:y] else matrix[i-1][j-1] + 0) # Adjusted the cost accordinly, insertion = 1, deletion=1 and substitution=2 return matrix[len_first][len_second] # Returning the final import scipy.stats as stats def kt(left, right): tau, p_value = stats.kendalltau(left, right) return tau def normalised_kendall_tau_distance(values1, values2): """Compute the Kendall tau distance.""" n = len(values1) assert len(values2) == n, "Both lists have to be of equal length" i, j = np.meshgrid(np.arange(n), np.arange(n)) a = np.argsort(values1) b = np.argsort(values2) ndisordered = np.logical_or(np.logical_and(a[i] < a[j], b[i] > b[j]), np.logical_and(a[i] > a[j], b[i] < b[j])).sum() return ndisordered / (n * (n - 1)) def jaccard_similarity(list1, list2): intersection = len(list(set(list1).intersection(list2))) union = (len(set(list1)) + len(set(list2))) - intersection return float(intersection) / union def distance_between_arrays(left, right): if len(left) != len(right): print('Arrays sizes does not match') jaccard = jaccard_similarity(left, right) return jaccard

countries_dict = get_all_countries_tracks_dict() #for country in countries_dict.keys(): # print(country, countries_dict[country])

def normalised_kendall_tau_distance(values1, values2): """Compute the Kendall tau distance.""" n = len(values1) assert len(values2) == n, "Both lists have to be of equal length" i, j = np.meshgrid(np.arange(n), np.arange(n)) a = np.argsort(values1) b = np.argsort(values2) ndisordered = np.logical_or(np.logical_and(a[i] < a[j], b[i] > b[j]), np.logical_and(a[i] > a[j], b[i] < b[j])).sum() return ndisordered / (n * (n - 1))

corr_data = [] for country1 in list(countries_dict.keys()): for country2 in list(countries_dict.keys()): d = distance_between_arrays(countries_dict[country1], countries_dict[country2]) corr_data.append([country1, country2, d])

corr_data = [[x[0],x[1], x[2]] for x in corr_data]

import networkx as nx import matplotlib.pyplot as plt from pyvis.network import Network G=nx.Graph() existing_nodes = [] existing_edges = [] def build_graph_for_all(): count=0 for d in corr_data: if d[0] not in existing_edges : G.add_node(str(d[0])) existing_nodes.append(str(d[0])) if d[1] not in existing_edges : G.add_node(str(d[1])) existing_nodes.append(str(d[1])) if d[0] != d[1] and not (((d[0], d[1]) in existing_edges) or ((d[1], d[0]) in existing_edges)): existing_edges.append((d[0], d[1])) G.add_weighted_edges_from([[str(d[0]), str(d[1]), d[2]]]) #build_graph_for_all() #pos = nx.spring_layout(G, weight='weight') #nx.draw(G, pos=pos, width=1, with_labels=True) #plt.savefig("path1.png")

cor_array = [] columns = ['Argentina', 'Chile', 'US', 'Spain', 'UK'] for i in range(0, len(corr_data), 5): tmp = [] for c in corr_data[i : i+5]: tmp.append(c[2]) cor_array.append(tmp)

from string import ascii_letters import numpy as np import pandas as pd import seaborn as sns import matplotlib.pyplot as plt d = pd.DataFrame(data=cor_array, columns=columns) # Compute the correlation matrix corr = d.corr() # Generate a mask for the upper triangle mask = np.triu(np.ones_like(corr, dtype=bool)) # Set up the matplotlib figure f, ax = plt.subplots(figsize=(11, 9)) # Generate a custom diverging colormap # Create a custom divergin palette cmap = sns.diverging_palette(100, 7, s=75, l=40, n=5, center="light", as_cmap=True) # Draw the heatmap with the mask and correct aspect ratio #sns.heatmap(corr, mask=mask, cmap=cmap, vmax=.3, center=0, # square=True, linewidths=.5, cbar_kws={"shrink": .5}) sns.heatmap(corr, mask=mask, center=0, annot=True, fmt='.2f', square=True, cmap='YlGnBu') ax.set_title('Jaccard Correlation Spotify Hits per Country') plt.show()