Un'analisi dell'attrattività turistica dell'Italia¶

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
%pylab inline

import Quandl as Q
TOKEN = '*****' # Da sostituire con il TOKEN ottenuto alla registrazione su Quandl

import re
import urllib2

def get_quandl_country_code(url, pattern):
    """ Get the country code from an URL and a pattern.
    
    Suppose your pattern have the 'country' group name.
    """
    rawtext = urllib2.urlopen(url).read()
    regexp = re.compile(pattern)
    return [m.group('country') for m in regexp.finditer(rawtext)]

URL = 'http://www.quandl.com/society/international-tourism-arrivals-all-countries'
PATTERN = r'WORLDBANK/(?P<country>[A-Z]+)_ST_INT_ARVL'
country_code = get_quandl_country_code(URL, PATTERN)[:20] # [:20] --> Paesi del G20
print 'Nb countries', len(country_code), ': ', country_code

def get_quandl_dataset(code, name=None):
    """ Get a Pandas DataFrame from the Quandl website according to the code.
    
    name: str or list
        The name(s) of the col label.
    """
    df = Q.get(code, authtoken=TOKEN)
    if name:
        name = [name] if isinstance(name, str) else name
        df.columns = pd.Index(name)
    return df

code = PATTERN.replace(r"(?P<country>[A-Z]+)", "%s")
inbound = pd.concat([get_quandl_dataset(code % name, name) for name in country_code], axis=1)

inbound.loc['2011-12-31', :].order(ascending=False)

def bar_chart(series, title='', xlabel= '', ylabel=''):
    """ Plot a bar chart from a Pandas series. 
    
    title: str
        The title of the chart.
    xlabel: str
        The label of x-axis.
    ylabel: str
        The label of y-axis.
    """
    fig = plt.figure(figsize=(10,10))
    ax = fig.add_axes([0.1, 0.1, 0.8, 0.8])
    ax.set_title(title)
    x = np.arange(len(series.values))
    ax.bar(x, series.values)
    ax.set_xticks(x + 0.4)
    ax.set_xticklabels(series.index)
    ax.set_xlabel(xlabel)
    ax.set_ylabel(ylabel)
    plt.show()

bar_chart(inbound.loc['2011-12-31', :].order(ascending=False), title='Arrivi turistici internazionali 2011', ylabel='Decine milioni')

(inbound.loc['2011-12-31', :] / inbound.loc['1995-12-31', :] - 1).order(ascending=False)

bar_chart((inbound.loc['2011-12-31', :] / inbound.loc['1995-12-31', :] - 1).order(ascending=False), title='Arrivi turistici internazionali: variazione 1995 - 2011', ylabel='Variazione assoluta')

URL2 = 'http://www.quandl.com/society/international-tourism-receipts-all-countries'
PATTERN2 = r'WORLDBANK/(?P<country>[A-Z]+)_ST_INT_RCPT_XP_ZS'
country_code = get_quandl_country_code(URL2, PATTERN2)[:20] # [:20] --> Paesi del G20
print 'Nb countries', len(country_code), ': ', country_code

code = PATTERN2.replace(r"(?P<country>[A-Z]+)", "%s")
inrecpt = pd.concat([get_quandl_dataset(code % name, name) for name in country_code], axis=1)

inrecpt.loc['2011-12-31', :].order(ascending=False)

bar_chart(inrecpt.loc['2011-12-31', :].order(ascending=False), title='Peso entrate turistiche internazionali su esportazioni', ylabel='Percentuale')

whc_dict = {'USA': 21, 'CHN': 45, 'JPN': 17, 'DEU': 38, 'FRA': 38, 'BRA': 19, 'GBR': 28, 'ITA': 49, 'RUS': 25, 'IND': 30, 'CAN': 17, 'AUS': 19, 'ESP': 44, 'MEX': 32, 'KOR': 10, 'IDN': 8, 'TUR': 11, 'SAU': 2, 'ARG': 8, 'ZAF': 8}
whc = pd.DataFrame.from_dict(whc_dict, orient='index')
whc.columns = ['WHC']
whc['WHC'].order(ascending=False)

bar_chart(whc['WHC'].order(ascending=False), title='Siti World Heritage per nazione', ylabel='# siti')

whc['Inbound'] = inbound.loc['2011-12-31', :]
whc['Ingrowth'] = (inbound.loc['2011-12-31', :] / inbound.loc['1995-12-31', :] - 1)
whc['Inreceipts'] = inrecpt.loc['2011-12-31', :]
whc

def bubble_chart(df, x=0, y=1, r=2, c=3, title='', xlabel='', ylabel=''):
    """ Plot a bubble chart.
    
    x: int
        Column of df for x data.
    y: int
        Column of df for y data.
    r: int
        Column of df for area data.
    c: int
        Column of df for color data.
    title: str
        The title of the chart.
    xlabel: str
        The label of x-axis.
    ylabel: str
        The label of y-axis.
    """
    fig = plt.figure(figsize=(10,10))
    ax = fig.add_axes([0.1, 0.1, 0.8, 0.8])
    ax.set_title(title)
    dx = df.iloc[:, x]
    dy = df.iloc[:, y]
    dr = df.iloc[:, r] * 2000 # La costante aggiusta l'area delle bolle in modo che sia chiaramente visibile
    dc = np.log(df.iloc[:, c]) # Il logaritmo serve per modificare la scala di colori riducendo le tonalità scure
    sct = ax.scatter(dx, dy, c=dc, s=dr)
    sct.set_alpha(0.3)
    for t in df.index:
        ax.text(df.loc[t, :].iloc[x], df.loc[t, :].iloc[y], t, ha='center', va='center')
    ax.set_xlabel(xlabel)
    ax.set_ylabel(ylabel)
    plt.show()    
    

bubble_chart(whc, title='Analisi flussi turistici internazionali', xlabel='# Siti World Heritage Unesco', ylabel='Arrivi internazionali 2011 (decine milioni)')