def calculate_diversity(data):
##################################################
### this part calculates user diversity
##################################################
div = defaultdict(int)
# for each user calculate diversity using formula from Finger on the Pulse
for usr in range(500001):
# total places visits
S = float(sum(data[usr].values()))
if S > 0:
for subpref in data[usr].iterkeys():
if subpref <> -1:
# frequency of visits to this place
v = data[usr][subpref] / float(S)
div[usr] -= v * math.log(v)
div[usr] = div[usr] / math.log(S)
# file_name = "/home/sscepano/D4D res/allstuff/diversity of travel/usr_diversity_index.tsv"
# f = open(file_name,"w")
#
# for usr in range(1,500001):
# f.write(str(usr) + '\t' + str(div[usr]) + '\n')
#
# print file_name
# f.close()
################################################################
### subpref diversity
################################################################
num_subpref_usrs = rd.read_in_subpref_num_users()
div_subpref = defaultdict(int)
for subpref in range(1,256):
# for places where we have users
if num_subpref_usrs[subpref] > 0:
# read ids of those users
subpref_usrs = rd.read_in_subpref_users(subpref)
# for each user check if he belongs here (not the fastest way but still fine)
for usr in range(500001):
# total places visits
if subpref_usrs[usr] == 1:
# we sum all users diversity indexes and then avg
div_subpref[subpref] += div[usr]
div_subpref[subpref] = div_subpref[subpref] / float(num_subpref_usrs[subpref])
file_name2 = "/home/sscepano/D4D res/allstuff/diversity of travel/subpref_diversity_index.tsv"
f2 = open(file_name2,"w")
for subpref in range(1,256):
if num_subpref_usrs[subpref] > 0:
f2.write(str(subpref) + '\t' + str(div_subpref[subpref]) + '\n')
print file_name2
f2.close()
return div, div_subpref
def from_file_radius_gyr3(file_name, subpref):
users_list = rd.read_in_subpref_users(subpref)
total = float(rd.read_in_subpref_num_users()[subpref])
if total > 0:
nits = []
its = []
# a loop where we populate those two arrays from the file
i = 0
f = open(file_name, 'r')
# read the file
for line in f:
i = i + 1
it, nit = line.split('\t')
nit = float(nit)
it = int(it)
if users_list[it] == 1:
nit = int(nit)
nits.append(nit)
its.append(it)
mi = min(nits)
mx = max(nits)
print("Minimum radius of gyr ", mi)
print("Maximum radius of gyr ", mx)
total_nit = float(sum(nits))
print("Total radius of gyr ", total_nit)
pdf_nits = defaultdict(int)
for j in range(0, len(nits)):
pdf_nits[nits[j]] += 1
ordered = OrderedDict(sorted(pdf_nits.items(), key=lambda t: t[0]))
nits7s = []
its7s = []
test = 0
#total = 500000.0
for j in ordered.iterkeys():
nits7s.append(ordered[j]/total)
test += ordered[j]/total
its7s.append(j)
print test
############################################################################################################################
# THIS is to plot number of users pdf
############################################################################################################################
plt.figure(7)
plt.plot(its7s, nits7s, 'o', linewidth=0.5, label= 'distribution of Rg')
plt.xlabel('rg [km]')
plt.ylabel('P(rg)')
plt.legend()
# this is if we want loglog lot, otheriwse comment and uncomment next line for regular plot file
plt.yscale('log')
plt.xscale('log')
figure_name = "/home/sscepano/D4D res/allstuff/rg/1/rg_" + str(subpref) + ".png"
print(figure_name)
plt.savefig(figure_name, format = "png", dpi=300)
plt.clf()
return