def _build_user_graph(name):
if name.endswith("lang"):
name = name[: -len("lang")]
dist = stats.get_language_data(name)
return charts.PieChart(dist.items(), max_options=8)
elif name == "dropout":
data = stats.get_dropout_figures()
approx_data = stats.approximate(data)
chart = charts.MultiLineChart(approx_data, data_name="histogram", x_axis=(0.4, 1.0, 0.1), y_axis=(0, 1.0, 0.1))
chart.add_data("raw", data)
return chart
elif name == "prepostdiff":
data = [r["pre_post_diff"] for r in stats.get_global_rater_stats() if r["n_tests"] > 2 and r["pre_post_diff"]]
hist_data = stats.histogram(data, n_bins=11, normalize=False, x_min=-0.7, x_max=0.7)
chart = charts.LineChart(hist_data, data_name="histogram", x_axis=(-0.8, 0.8, 0.2))
chart.add_data("raw", data)
return chart
elif name == "abilityjlpt3":
data = stats.get_user_scores("jlpt 3")
hist_data = stats.histogram(data, x_min=0.0, x_max=1.0, normalize=False)
chart = charts.LineChart(hist_data, data_name="histogram", x_axis=(0.0, 1.0, 0.1))
chart.add_data("raw", data)
return chart
elif name == "abilityjlpt4":
data = stats.get_user_scores("jlpt 4")
hist_data = stats.histogram(data, x_min=0.0, x_max=1.0, normalize=False)
chart = charts.LineChart(hist_data, data_name="histogram", x_axis=(0.0, 1.0, 0.1))
chart.add_data("raw", data)
return chart
raise KeyError(name)
def _build_user_graph(name):
if name.endswith('lang'):
name = name[:-len('lang')]
dist = stats.get_language_data(name)
return charts.PieChart(dist.items(), max_options=8)
elif name == 'dropout':
data = stats.get_dropout_figures()
approx_data = stats.approximate(data)
chart = charts.MultiLineChart(approx_data,
data_name='histogram',
x_axis=(0.4, 1.0, 0.1),
y_axis=(0, 1.0, 0.1))
chart.add_data('raw', data)
return chart
elif name == 'prepostdiff':
data = [
r['pre_post_diff'] for r in stats.get_global_rater_stats()
if r['n_tests'] > 2 and r['pre_post_diff']
]
hist_data = stats.histogram(data,
n_bins=11,
normalize=False,
x_min=-0.7,
x_max=0.7)
chart = charts.LineChart(hist_data,
data_name='histogram',
x_axis=(-0.8, 0.8, 0.2))
chart.add_data('raw', data)
return chart
elif name == 'abilityjlpt3':
data = stats.get_user_scores('jlpt 3')
hist_data = stats.histogram(data,
x_min=0.0,
x_max=1.0,
normalize=False)
chart = charts.LineChart(hist_data,
data_name='histogram',
x_axis=(0.0, 1.0, 0.1))
chart.add_data('raw', data)
return chart
elif name == 'abilityjlpt4':
data = stats.get_user_scores('jlpt 4')
hist_data = stats.histogram(data,
x_min=0.0,
x_max=1.0,
normalize=False)
chart = charts.LineChart(hist_data,
data_name='histogram',
x_axis=(0.0, 1.0, 0.1))
chart.add_data('raw', data)
return chart
raise KeyError(name)
def statsex(self, objects):
"""
Do some statistics on a source list
Return dictionary
"""
import stats, pstat
# Return if we have no objects
if len(objects) == 0:
return 0
# Define dictionary to hold statistics
stat = {}
# Get number of objects
stat['N'] = str(len(objects))
# Define list (float) of FWHM values
fwhm = [ float(obj[7]) for obj in objects ]
# Define list (float) of ELLIPTICITY values
el = [ float(obj[6]) for obj in objects ]
# Define list (float) of THETA_IMAGE values
pa = [ float(obj[5]) for obj in objects ]
# Define list (float) of 'Stella-like' values
stella = [ float(obj[9]) for obj in objects ]
# Create a histogram of FWHM values of binsize 1 pixel
hfwhm = stats.histogram(fwhm,40,[0,40])[0]
stat['medianFWHM'] = "%.2f" % stats.median(fwhm)
stat['meanFWHM'] = "%.2f" % stats.mean(fwhm)
stat['modeFWHM'] = "%.2f" % float(hfwhm.index(max(hfwhm))+0.5)
try:
stat['stdevFWHM'] = "%.2f" % stats.stdev(fwhm)
except ZeroDivisionError:
stat['stdevFWHM'] = '0.00';
stat['medianEL'] = "%.2f" % stats.median(el)
stat['meanEL'] = "%.2f" % stats.mean(el)
try:
stat['stdevEL'] = "%.2f" % stats.stdev(el)
except ZeroDivisionError:
stat['stdevEL'] = '0.00'
# Histogram of Ellipticity PA (-180 to 180, bins of 45 deg)
#stat['histoTHETA'] = stats.histogram(pa,8,[-180,180])[0]
# Histogram of Stellarity (0 to 1, bins of 0.05)
#stat['histoStella'] = stats.histogram(stella,20,[0,1.01])[0]
return stat
def statsex(self, objects):
"""
Do some statistics on a source list
Return dictionary
"""
import stats, pstat
# Return if we have no objects
if len(objects) == 0:
return 0
# Define dictionary to hold statistics
stat = {}
# Get number of objects
stat['N'] = str(len(objects))
# Define list (float) of FWHM values
fwhm = [ float(obj[7]) for obj in objects ]
# Define list (float) of ELLIPTICITY values
el = [ float(obj[6]) for obj in objects ]
# Define list (float) of THETA_IMAGE values
pa = [ float(obj[5]) for obj in objects ]
# Define list (float) of 'Stella-like' values
stella = [ float(obj[9]) for obj in objects ]
# Create a histogram of FWHM values of binsize 1 pixel
hfwhm = stats.histogram(fwhm,40,[0,40])[0]
stat['medianFWHM'] = "%.2f" % stats.median(fwhm)
stat['meanFWHM'] = "%.2f" % stats.mean(fwhm)
stat['modeFWHM'] = "%.2f" % float(hfwhm.index(max(hfwhm))+0.5)
try:
stat['stdevFWHM'] = "%.2f" % stats.stdev(fwhm)
except ZeroDivisionError:
stat['stdevFWHM'] = '0.00';
stat['medianEL'] = "%.2f" % stats.median(el)
stat['meanEL'] = "%.2f" % stats.mean(el)
try:
stat['stdevEL'] = "%.2f" % stats.stdev(el)
except ZeroDivisionError:
stat['stdevEL'] = '0.00'
# Histogram of Ellipticity PA (-180 to 180, bins of 45 deg)
#stat['histoTHETA'] = stats.histogram(pa,8,[-180,180])[0]
# Histogram of Stellarity (0 to 1, bins of 0.05)
#stat['histoStella'] = stats.histogram(stella,20,[0,1.01])[0]
return stat
def _build_test_graph(name):
if name == 'mean':
score_data = stats.get_mean_score_nth_test()
data = stats.group_by_points(score_data, y_max=1.0, y_min=0.0)
chart = charts.MultiLineChart(data,
y_axis=(0, 1, 0.1),
data_name='grouped')
chart.add_data('raw', score_data)
return chart
elif name == 'volume':
user_data = stats.get_users_by_n_tests()
return charts.LineChart(user_data)
elif name == 'length':
return charts.PieChart(stats.get_test_length_volume())
elif name == 'normtime':
user_data = stats.get_score_over_norm_time()
return charts.LineChart(user_data)
elif name == 'time':
base_data = stats.get_score_over_time()
data = stats.approximate(base_data)
chart = charts.MultiLineChart(data,
y_axis=(0, 1.05, 0.1),
data_name='approximate')
chart.add_data('raw', base_data)
two_colours = charts.color_desc(2).split(',')
three_colours = ','.join(
(two_colours[0], two_colours[1], two_colours[1]))
chart['chco'] = three_colours
return chart
elif name == 'dropout':
return charts.LineChart(stats.get_mean_score_by_n_tests())
elif name == 'firstlast':
data = stats.get_first_last_test()
hist_data = stats.histogram(data,
n_bins=11,
normalize=False,
x_min=-0.5,
x_max=0.5)
chart = charts.LineChart(hist_data,
data_name='histogram',
x_axis=(-0.5, 0.5, 0.1))
chart.add_data('raw', data)
return chart
raise KeyError(name)
def hist(x, nbins=15):
"""Simple histogram function.
**x** -- data set as numeric vector
**nbins** -- number of histogram bins
matlab equiv: HIST
"""
import stats
counts, smallest, binsize, extras = stats.histogram(x, nbins)
graceplot().histoPlot(counts,
x_min=smallest,
x_max=smallest+len(counts)*binsize,
fillcolor=2, edgecolor=1, labeled=0)
def get_sentiword_score(self):
f = open("SentiWordNet/data/SentiWord.json", "r")
self.sentiwords = cjson.decode(f.readline())
f.close()
self.read_user_vector("user_vector/user_vector_new.json")
self.neg_words = self.get_word_list(
"SentiWordNet/data/negative_words_list.txt")
self.pos_words = self.get_word_list(
"SentiWordNet/data/positive_words_list.txt")
myht = histogram()
for user in self.user_vector:
for tweet in self.user_vector[user]["tweets"]:
words = self.process_sentence(tweet["text"])
pnword_score = 0.0
sentiword_score = 0.0
sentiword_count = 0
for word in words:
if self.sentiwords.has_key(word):
sentiword_score += self.sentiwords[word]
sentiword_count += 1
if self.pos_words.has_key(word):
pnword_score += 1.0
elif self.neg_words.has_key(word):
pnword_score -= 1.0
myht.add(pnword_score)
tweet.update({"pnword_score": pnword_score})
if sentiword_count != 0:
sentiword_score = sentiword_score / float(sentiword_count)
tweet.update({"sentiword_score": sentiword_score})
#print self.user_vector[user]["tweets"]
mean = myht.avg()
std = myht.std()
# normalize the positive_negative_word score
for user in self.user_vector:
for tweet in self.user_vector[user]["tweets"]:
tweet["pnword_score"] = (tweet["pnword_score"] - mean) / std
#print tweet["pnword_score"]
f = open("user_vector/user_vector_new_2.json", "w")
for user in self.user_vector:
json.dump(self.user_vector[user], f)
f.write("\n")
f.close()
def hist(x, nbins=15):
"""Simple histogram function.
**x** -- data set as numeric vector
**nbins** -- number of histogram bins
matlab equiv: HIST
"""
import stats
counts, smallest, binsize, extras = stats.histogram(x, nbins)
graceplot().histoPlot(counts,
x_min=smallest,
x_max=smallest + len(counts) * binsize,
fillcolor=2,
edgecolor=1,
labeled=0)
def get_sentiword_score(self):
f = open("SentiWordNet/data/SentiWord.json","r")
self.sentiwords = cjson.decode(f.readline())
f.close()
self.read_user_vector("user_vector/user_vector_new.json")
self.neg_words = self.get_word_list("SentiWordNet/data/negative_words_list.txt")
self.pos_words = self.get_word_list("SentiWordNet/data/positive_words_list.txt")
myht = histogram()
for user in self.user_vector:
for tweet in self.user_vector[user]["tweets"]:
words = self.process_sentence(tweet["text"])
pnword_score = 0.0
sentiword_score = 0.0
sentiword_count = 0
for word in words:
if self.sentiwords.has_key(word):
sentiword_score += self.sentiwords[word]
sentiword_count += 1
if self.pos_words.has_key(word):
pnword_score += 1.0
elif self.neg_words.has_key(word):
pnword_score -= 1.0
myht.add(pnword_score)
tweet.update({"pnword_score" : pnword_score})
if sentiword_count != 0:
sentiword_score = sentiword_score/float(sentiword_count)
tweet.update({"sentiword_score" : sentiword_score})
#print self.user_vector[user]["tweets"]
mean = myht.avg()
std = myht.std()
# normalize the positive_negative_word score
for user in self.user_vector:
for tweet in self.user_vector[user]["tweets"]:
tweet["pnword_score"] = (tweet["pnword_score"]-mean)/std
#print tweet["pnword_score"]
f = open("user_vector/user_vector_new_2.json", "w")
for user in self.user_vector:
json.dump(self.user_vector[user], f)
f.write("\n")
f.close()
def _build_test_graph(name):
if name == "mean":
score_data = stats.get_mean_score_nth_test()
data = stats.group_by_points(score_data, y_max=1.0, y_min=0.0)
chart = charts.MultiLineChart(data, y_axis=(0, 1, 0.1), data_name="grouped")
chart.add_data("raw", score_data)
return chart
elif name == "volume":
user_data = stats.get_users_by_n_tests()
return charts.LineChart(user_data)
elif name == "length":
return charts.PieChart(stats.get_test_length_volume())
elif name == "normtime":
user_data = stats.get_score_over_norm_time()
return charts.LineChart(user_data)
elif name == "time":
base_data = stats.get_score_over_time()
data = stats.approximate(base_data)
chart = charts.MultiLineChart(data, y_axis=(0, 1.05, 0.1), data_name="approximate")
chart.add_data("raw", base_data)
two_colours = charts.color_desc(2).split(",")
three_colours = ",".join((two_colours[0], two_colours[1], two_colours[1]))
chart["chco"] = three_colours
return chart
elif name == "dropout":
return charts.LineChart(stats.get_mean_score_by_n_tests())
elif name == "firstlast":
data = stats.get_first_last_test()
hist_data = stats.histogram(data, n_bins=11, normalize=False, x_min=-0.5, x_max=0.5)
chart = charts.LineChart(hist_data, data_name="histogram", x_axis=(-0.5, 0.5, 0.1))
chart.add_data("raw", data)
return chart
raise KeyError(name)
print('var:',stats.var(a),stats.var(af))
print('stdev:',stats.stdev(a),stats.stdev(af))
print('sem:',stats.sem(a),stats.sem(af))
print('describe:')
print(stats.describe(l))
print(stats.describe(lf))
print(stats.describe(a))
print(stats.describe(af))
print('\nFREQUENCY')
print('freqtable:')
print('itemfreq:')
print(stats.itemfreq(l))
print(stats.itemfreq(a))
print('scoreatpercentile:',stats.scoreatpercentile(l,40),stats.scoreatpercentile(lf,40),stats.scoreatpercentile(a,40),stats.scoreatpercentile(af,40))
print('percentileofscore:',stats.percentileofscore(l,12),stats.percentileofscore(lf,12),stats.percentileofscore(a,12),stats.percentileofscore(af,12))
print('histogram:',stats.histogram(l),stats.histogram(a))
print('cumfreq:')
print(stats.cumfreq(l))
print(stats.cumfreq(lf))
print(stats.cumfreq(a))
print(stats.cumfreq(af))
print('relfreq:')
print(stats.relfreq(l))
print(stats.relfreq(lf))
print(stats.relfreq(a))
print(stats.relfreq(af))
print('\nVARIATION')
print('obrientransform:')
l = range(1,21)
a = N.array(l)
ll = [l]*5
def draw_distributions(self):
"""
draw distributions for all terms in self.terms and save figs to specified folders
"""
f = open("topics/term_senti_scores.json", "r")
self.terms = cjson.decode(f.readline())
f.close()
print len(self.terms)
count = 0
for term in self.terms:
hsw = stats.histogram()
#hpn = stats.histogram()
hst = stats.histogram()
has = stats.histogram()
for s in self.terms[term]["sentiword_score"]:
hsw.add(s)
#for s in self.terms[term]["pnword_score"]:
# hpn.add(s)
for s in self.terms[term]["sentiment_score"]:
hst.add(s)
for s in self.terms[term]["avg_sentiscore"]:
has.add(s)
distribution_sw = hsw.histogram_2()
#distribution_pn = hpn.histogram_2()
distribution_st = hst.histogram_2()
distribution_as = has.histogram_2()
# sentiword score
x_axis = range(len(distribution_sw))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i * 0.05 for i in range(42)]
plt.bar(x_axis,
distribution_sw,
width=0.8,
facecolor='blue',
alpha=0.5)
X_ticks = ["%.1f" % (x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i - 1) % 2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Sentiword Score")
plt.ylabel("Percentage (%)")
plt.title("Sentiword Score Distribution - %s" % term)
plt.savefig(sentiword_folder + "%s.png" % term, dpi=50)
plt.clf()
"""
# positive negative score
x_axis = range(len(distribution_pn))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i*0.05 for i in range(42)]
plt.bar(x_axis, distribution_pn, width=0.8, facecolor='blue', alpha = 0.5)
X_ticks = ["%.1f" %(x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i-1)%2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Positive-Negative Word Score")
plt.ylabel("Percentage (%)")
plt.title("PNWord Score Distribution - %s" %term)
plt.savefig(pnword_folder+"%s.png" %term,dpi=50)
plt.clf()
"""
# sentiment score
x_axis = range(len(distribution_st))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i * 0.05 for i in range(42)]
plt.bar(x_axis,
distribution_st,
width=0.8,
facecolor='blue',
alpha=0.5)
X_ticks = ["%.1f" % (x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i - 1) % 2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Sentiment Score")
plt.ylabel("Percentage (%)")
plt.title("Sentiment Score Distribution - %s" % term)
plt.savefig(sentiment_folder + "%s.png" % term, dpi=50)
plt.clf()
# sentiscore score
x_axis = range(len(distribution_as))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i * 0.05 for i in range(42)]
plt.bar(x_axis,
distribution_as,
width=0.8,
facecolor='blue',
alpha=0.5)
X_ticks = ["%.1f" % (x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i - 1) % 2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Avg Senti Score")
plt.ylabel("Percentage (%)")
plt.title("Senti Score Distribution - %s" % term)
plt.savefig(sentiscore_folder + "%s.png" % term, dpi=50)
plt.clf()
count += 1
print "done %s: neutral_st %d neutral_sw %d neutral_as %d count %d" % (
term, hst.count_zero(), hsw.count_zero(), has.count_zero(),
count)
print 'tstdev:',stats.tstdev(a,(5,17)),stats.tstdev(af,(5,17)) print 'tsem:',stats.tsem(a,(5,17)),stats.tsem(af,(5,17)) print 'describe:' print stats.describe(l) print stats.describe(lf) print stats.describe(a) print stats.describe(af) print '\nFREQUENCY' print 'freqtable:' print 'itemfreq:' print stats.itemfreq(l) print stats.itemfreq(a) print 'scoreatpercentile:',stats.scoreatpercentile(l,40),stats.scoreatpercentile(lf,40),stats.scoreatpercentile(a,40),stats.scoreatpercentile(af,40) print 'percentileofscore:',stats.percentileofscore(l,12),stats.percentileofscore(lf,12),stats.percentileofscore(a,12),stats.percentileofscore(af,12) print 'histogram:',stats.histogram(l),stats.histogram(a) print 'cumfreq:' print stats.cumfreq(l) print stats.cumfreq(lf) print stats.cumfreq(a) print stats.cumfreq(af) print 'relfreq:' print stats.relfreq(l) print stats.relfreq(lf) print stats.relfreq(a) print stats.relfreq(af) print '\nVARIATION' print 'obrientransform:' l = range(1,21)
def draw_distributions(self):
"""
draw distributions for all terms in self.terms and save figs to specified folders
"""
f = open("topics/term_senti_scores.json","r")
self.terms = cjson.decode(f.readline())
f.close()
print len(self.terms)
count = 0
for term in self.terms:
hsw = stats.histogram()
#hpn = stats.histogram()
hst = stats.histogram()
has = stats.histogram()
for s in self.terms[term]["sentiword_score"]:
hsw.add(s)
#for s in self.terms[term]["pnword_score"]:
# hpn.add(s)
for s in self.terms[term]["sentiment_score"]:
hst.add(s)
for s in self.terms[term]["avg_sentiscore"]:
has.add(s)
distribution_sw = hsw.histogram_2()
#distribution_pn = hpn.histogram_2()
distribution_st = hst.histogram_2()
distribution_as = has.histogram_2()
# sentiword score
x_axis = range(len(distribution_sw))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i*0.05 for i in range(42)]
plt.bar(x_axis, distribution_sw, width=0.8, facecolor='blue', alpha = 0.5)
X_ticks = ["%.1f" %(x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i-1)%2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Sentiword Score")
plt.ylabel("Percentage (%)")
plt.title("Sentiword Score Distribution - %s" %term)
plt.savefig(sentiword_folder+"%s.png" %term,dpi=50)
plt.clf()
"""
# positive negative score
x_axis = range(len(distribution_pn))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i*0.05 for i in range(42)]
plt.bar(x_axis, distribution_pn, width=0.8, facecolor='blue', alpha = 0.5)
X_ticks = ["%.1f" %(x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i-1)%2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Positive-Negative Word Score")
plt.ylabel("Percentage (%)")
plt.title("PNWord Score Distribution - %s" %term)
plt.savefig(pnword_folder+"%s.png" %term,dpi=50)
plt.clf()
"""
# sentiment score
x_axis = range(len(distribution_st))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i*0.05 for i in range(42)]
plt.bar(x_axis, distribution_st, width=0.8, facecolor='blue', alpha = 0.5)
X_ticks = ["%.1f" %(x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i-1)%2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Sentiment Score")
plt.ylabel("Percentage (%)")
plt.title("Sentiment Score Distribution - %s" %term)
plt.savefig(sentiment_folder+"%s.png" %term,dpi=50)
plt.clf()
# sentiscore score
x_axis = range(len(distribution_as))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i*0.05 for i in range(42)]
plt.bar(x_axis, distribution_as, width=0.8, facecolor='blue', alpha = 0.5)
X_ticks = ["%.1f" %(x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i-1)%2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Avg Senti Score")
plt.ylabel("Percentage (%)")
plt.title("Senti Score Distribution - %s" %term)
plt.savefig(sentiscore_folder+"%s.png" %term, dpi=50)
plt.clf()
count += 1
print "done %s: neutral_st %d neutral_sw %d neutral_as %d count %d" %(term, hst.count_zero(), hsw.count_zero(), has.count_zero(), count)
print stats.describe(lf)
print stats.describe(a)
print stats.describe(af)
print '\nFREQUENCY'
print 'freqtable:'
print 'itemfreq:'
print stats.itemfreq(l)
print stats.itemfreq(a)
print 'scoreatpercentile:', stats.scoreatpercentile(
l, 40), stats.scoreatpercentile(lf, 40), stats.scoreatpercentile(
a, 40), stats.scoreatpercentile(af, 40)
print 'percentileofscore:', stats.percentileofscore(
l, 12), stats.percentileofscore(lf, 12), stats.percentileofscore(
a, 12), stats.percentileofscore(af, 12)
print 'histogram:', stats.histogram(l), stats.histogram(a)
print 'cumfreq:'
print stats.cumfreq(l)
print stats.cumfreq(lf)
print stats.cumfreq(a)
print stats.cumfreq(af)
print 'relfreq:'
print stats.relfreq(l)
print stats.relfreq(lf)
print stats.relfreq(a)
print stats.relfreq(af)
print '\nVARIATION'
print 'obrientransform:'
l = range(1, 21)
def read_data_in_range(filename = "./", topschoolfile = "./", start_year = 2000, end_year = 2014, self_edge = True):
"""
@description: read the recent data back until specified <cutting_year>
@type filename: string
@param filename: input file path and name
@type start_year: integer
@param start_year: the earliest year to be considered
@type end_year: integer
@param end_year: the latest year to be considered
@type self_edge: Boolean
@param self_edge: whether self edges are included or not; True-yes, False-not
@return: list of nodes
@return: list of edges
"""
top_50 = []
f = open(topschoolfile,"r")
for line in f:
line = line.strip().lower()
top_50.append(line)
f.close()
s = {}
edge_list_all = []
f = open(filename,"r")
f.readline() # skip the first row
for line in f:
line = line.lower()
line = line.strip() # remove those "\r\n"
lines = line.split(",") ## subject to change
if len(lines) == 2 or len(lines) == 3:
# if lines[0].strip() in top_50 and lines[1].strip() in top_50:
edge = []
for i in range(2):
edge.append(lines[i].strip())
if s.has_key(lines[i].strip()):
s[lines[i].strip()] += 1
else:
s.update({lines[i].strip() : 1})
if len(lines) == 2: # without year data
edge.append("-") ## never enter this loop
else:
#print lines
if len(lines[2]) > 0: # with year data
edge.append(lines[2].strip())
else: # without year data
pass
edge_list_all.append(edge)
f.close()
## statistical analysis
hist = stats.histogram()
stat = {}
cnt = 0
## re-organize the edge with weights
edge_dict = {}
for edge in edge_list_all:
if len(edge) == 3 and int(edge[2]) >= start_year and int(edge[2]) <= end_year: ## filtering the recent faculty data
cnt += 1
key = edge[0]+"#"+edge[1]
hist.add(edge[2].strip())
if not stat.has_key(edge[0]):
stat.update({edge[0] : {'total' : 1, 'wyear' : 1}})
else:
stat[edge[0]]['total'] += 1
stat[edge[0]]['wyear'] += 1
if edge_dict.has_key(key):
edge_dict[key] += 1.0
else:
edge_dict.update({key : 1.0})
else:
if not stat.has_key(edge[0]):
stat.update({edge[0] : {'total' : 1, 'wyear' : 0}})
else:
stat[edge[0]]['total'] += 1
# # statistics
# index, dist, cdf = hist.cdf()
# print hist._max, hist._min
# print len(index), index
# print len(dist), dist
# print len(cdf), cdf
#
# f = open("../result/result_top50_cs_newdata_apr09/year_statistical_from%d_to%d_extended.csv" %(start_year, end_year),"w")
# f.write("univ,total,wyear\n")
# for key in stat:
# f.write("%s,%d,%d\n" %(key, stat[key]['total'], stat[key]['wyear']))
# f.close()
#
# # the CDF of year distribution
# f = open("../result/result_top50_cs_newdata_apr09/year_cdf_from%d_to%d_extended.csv" %(start_year, end_year),"w")
# f.write("year,freq,percentile\n")
# for i in range(len(index)):
# f.write("%s,%d,%.3f\n" %(index[i], int(dist[i]), cdf[i]))
# f.close()
edge_list = []
for item in edge_dict.iteritems():
edge = []
univs = item[0].split("#")
if not self_edge == True:
if not univs[0].strip() == univs[1].strip():
edge.append(univs[0].strip())
edge.append(univs[1].strip())
edge.append(item[1])
edge_list.append(edge)
else:
pass
else:
edge.append(univs[0].strip())
edge.append(univs[1].strip())
edge.append(item[1])
edge_list.append(edge)
#print len(edge_list), edge_list
node_list = sorted(s.keys(), reverse = False)
return node_list, edge_list
def keyword_distribution(self, rm_en = True):
f = open("results/statistics_hashtag_sentiscores_%d.csv" %len(self.keywords),"w")
f.write("keyword,user_count,tag_count,count,sw_avg_neg,sw_avg_pos,sw_count_neg,sw_count_zero,sw_count_pos,sw_min,sw_max,"
+"st_avg_neg,st_avg_pos,st_count_neg,st_count_zero,st_count_pos,st_min,st_max,"
+"sc_avg_neg,sc_avg_pos,sc_count_neg,sc_count_zero,sc_count_pos,sc_min,sc_max\n")
for word in self.keywords:
hsw = stats.histogram()
hst = stats.histogram()
has = stats.histogram()
for score in self.keywords[word]["scores"]:
hsw.add(score["sentiword_score"])
hst.add(score["sentiment_score"])
has.add(score["sentiscore"])
f.write("%s,%d,%d,%d,%.3f,%.3f,%d,%d,%d,%.3f,%.3f," %(word, self.keywords[word]["user_count"], self.keywords[word]["tag_count"], hsw._count,
hsw._mean_neg, hsw._mean_pos, hsw._count_neg, hsw._zero, hsw._count_pos, hsw._min, hsw._max))
f.write("%.3f,%.3f,%d,%d,%d,%.3f,%.3f," %(hst._mean_neg, hst._mean_pos, hst._count_neg, hst._zero, hst._count_pos, hst._min, hst._max))
f.write("%.3f,%.3f,%d,%d,%d,%.3f,%.3f\n" %(has._mean_neg, has._mean_pos, has._count_neg, has._zero, has._count_pos, has._min, has._max))
distribution_sw = hsw.histogram_2()
distribution_st = hst.histogram_2()
distribution_as = has.histogram_2()
# sentiword score
x_axis = range(len(distribution_sw))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i*0.05 for i in range(42)]
plt.bar(x_axis, distribution_sw, width=0.8, facecolor='blue', alpha = 0.5)
X_ticks = ["%.1f" %(x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i-1)%2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Sentiword Score")
plt.ylabel("Percentage (%)")
plt.title("Sentiword Score Distribution - %s" %word)
plt.savefig(OUT_DIR+SUB_DIR_1+"%s.png" %word,dpi=50)
plt.clf()
# sentiment score
x_axis = range(len(distribution_st))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i*0.05 for i in range(42)]
plt.bar(x_axis, distribution_st, width=0.8, facecolor='blue', alpha = 0.5)
X_ticks = ["%.1f" %(x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i-1)%2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Sentiment Score")
plt.ylabel("Percentage (%)")
plt.title("Sentiment Score Distribution - %s" %word)
plt.savefig(OUT_DIR+SUB_DIR_2+"%s.png" %word,dpi=50)
plt.clf()
# sentiscore score
x_axis = range(len(distribution_as))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i*0.05 for i in range(42)]
plt.bar(x_axis, distribution_as, width=0.8, facecolor='blue', alpha = 0.5)
X_ticks = ["%.1f" %(x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i-1)%2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Avg Senti Score")
plt.ylabel("Percentage (%)")
plt.title("Senti Score Distribution - %s" %word)
plt.savefig(OUT_DIR+SUB_DIR_3+"%s.png" %word, dpi=50)
plt.clf()
print "done %s: neutral_st %d neutral_sw %d neutral_as %d" %(word, hst.count_zero(), hsw.count_zero(), has.count_zero())
f.close()
'lensr_v1': (1.0, 3.0)
}
# Start the nestlefit with the hsiao-stretch or the custom_model
# Save the parameters in 'nestfitparam.dat' and plot the model
MI_mH = MI_model(mH, 4, f=1.7, useprior=True, samerv=True)
model, res = nest_lc(phot_d, MI_mH, fit_param, fit_bounds)
#MI_geu_source = MI_model(geu_source,4)
#model,res = nest_lc(phot_d,MI_geu_source,fit_param_16geu,fit_bounds)
# Draw the correlations between the parameters that are degenerated
samp, params = res.samples, res.vparam_names
samples(samp, params, fit_param[9:])
# Draw the probability density function
df, data = res.ndof, model.tot_amp
fmin, minamp = min(model.chi), data[np.argmin(model.chi)]
histogram(data)
# Plot the lightcurves for the model and write the parameters into a file
model.plot(phot_d)
myfile = open('nestfitparam.dat', 'w')
for name in fit_param:
myfile.write(name + ' = ' + str(model.get(name)) + ' (' +
str(res.errors[name]) + ')\n')
myfile.write("chiqsquare = " + str(fmin) + ' Dof = ' + str(df) + '\n')
myfile.write('total amplification = ' + str(minamp) + '\n')
myfile.write('Bayesian evidence z = ' + str(np.exp(res.logz)))
myfile.close()
import json
import operator
import pylab
import sys
from stats import histogram
from pylabutils import setupPlot
prev = None
A = []
for line in sys.stdin:
data = json.loads(line)
acc = data.get('accuracy')
if acc:
A.append(min(999, acc))
w = 10.0
XY = histogram(A, w)
X = map(operator.itemgetter(0), XY)
Y = map(operator.itemgetter(1), XY)
pylab.bar(X, Y, width=w)
setupPlot(
'Accuracy (m)',
'Frequency',
'Accuracy Radius'
)
pylab.savefig('accuracy-histogram.png')
def train(notations, output):
hist = stats.histogram(read_notations(notations))
yaml.dump(hist, output)
root = etree.parse(sys.stdin).getroot()
document = root.getchildren()[-1]
placemark = document.getchildren()[-1]
track = placemark.getchildren()[-1]
dts = []
prev_when = None
for child in track.iterchildren():
if child.tag.endswith('when'):
ts = dateutil.parser.parse(child.text)
when = int(time.mktime(ts.timetuple()))
if prev_when is not None:
dts.append(when - prev_when)
prev_when = when
rmv = RollingMeanVar(0.001)
Y = []
for i, dt in enumerate(dts):
rmv.update(dt, i)
Y.append(rmv.mean())
pylab.plot(range(len(Y)), Y, color=RED)
setupPlot('Sample Index', 'Interval (s)', 'Polling Interval - Rolling Average')
pylab.savefig('timings-frequency.png')
pylab.close()
Y = map(operator.itemgetter(1), histogram(dt % 60 for dt in dts))
pylab.bar(range(len(Y)), Y, width=1.0)
setupPlot('Interval (s, mod 60)', 'Frequency',
'Polling Interval - Seconds mod 60')
pylab.savefig('timings-second-histogram.png')
pylab.close()
print
print("ACCOUNTS")
print("--------------------------------------------------")
print("Number of accounts: %s" % (len(accounts)))
print
print("HOURS OF PLAY")
print("--------------------------------------------------")
print("Total play time: %s" % getHourString(stats.sum(table[TABLE_TIME])))
print("Largest play time for a single account: %s" %
getHourString(max(table[TABLE_TIME])))
print("Median total play time for all accounts: %s" %
getHourString(stats.median(table[TABLE_TIME])))
print(
"Macro Histogram play time: \n%s" %
getHistogramString(stats.histogram(table[TABLE_TIME], 10, [0, 100 * 3600]),
getHourString))
print(
"Micro Histogram play time: \n%s" % getHistogramString(
stats.histogram(table[TABLE_TIME], 10, [0, 10 * 3600]), getHourString))
print(
"Pico Histogram play time: \n%s" % getHistogramString(
stats.histogram(table[TABLE_TIME], 12, [0, 1 * 3600]), getHourString))
print
print("LOGINS")
print("--------------------------------------------------")
print("Total logins: %s" % stats.sum(table[TABLE_LOGINS]))
print("Largest logins for a single account: %s" % max(table[TABLE_LOGINS]))
print("Median number of logins for all accounts: %s" %
stats.median(table[TABLE_LOGINS]))
import json
import operator
import pylab
import sys
from stats import histogram
from pylabutils import setupPlot
prev = None
A = []
for line in sys.stdin:
data = json.loads(line)
acc = data.get('accuracy')
if acc:
A.append(min(999, acc))
w = 10.0
XY = histogram(A, w)
X = map(operator.itemgetter(0), XY)
Y = map(operator.itemgetter(1), XY)
pylab.bar(X, Y, width=w)
setupPlot('Accuracy (m)', 'Frequency', 'Accuracy Radius')
pylab.savefig('accuracy-histogram.png')
def read_data(filename, topschoolfile, self_edge = True, extended = True):
"""
@type filename: string
@param filename: input file path and name
@type self_edge: Boolean
@param self_edge: whether self edges are included or not; True-yes, False-not
@type extended: Boolean
@param extended: whether the graph is extended or restricted in top schools or not; default is True
@return: list of nodes
@return: list of edges
"""
top_50 = []
f = open(topschoolfile,"r")
for line in f:
line = line.strip().lower()
top_50.append(line)
f.close()
## statistical analysis
hist = stats.histogram()
stat = {}
s = {}
edge_list_all = []
f = open(filename,"r")
# print f.readline() # skip the first row
for line in f:
line = line.lower()
line = line.strip() # remove those "\r\n"
lines = line.split(",") ## subject to change
if len(lines) == 2 or len(lines) == 3:
if extended == True:
edge = []
for i in range(2):
edge.append(lines[i].strip())
if s.has_key(lines[i].strip()):
s[lines[i].strip()] += 1
else:
s.update({lines[i].strip() : 1})
if len(lines) == 2: # without year data
edge.append("-")
if not stat.has_key(lines[0]):
stat.update({lines[0] : {'total' : 1, 'wyear' : 0}})
else:
stat[lines[0]]['total'] += 1
else:
#print lines
if len(lines[2]) > 0: # with year data
edge.append(lines[2].strip())
hist.add(lines[2].strip())
if not stat.has_key(lines[0]):
stat.update({lines[0] : {'total' : 1, 'wyear' : 1}})
else:
stat[lines[0]]['total'] += 1
stat[lines[0]]['wyear'] += 1
else: # without year data
if not stat.has_key(lines[0]):
stat.update({lines[0] : {'total' : 1, 'wyear' : 0}})
else:
stat[lines[0]]['total'] += 1
edge_list_all.append(edge)
else:
if lines[0].strip() in top_50 and lines[1].strip() in top_50:
edge = []
for i in range(2):
edge.append(lines[i].strip())
if s.has_key(lines[i].strip()):
s[lines[i].strip()] += 1
else:
s.update({lines[i].strip() : 1})
if len(lines) == 2: # without year data
edge.append("-")
if not stat.has_key(lines[0]):
stat.update({lines[0] : {'total' : 1, 'wyear' : 0}})
else:
stat[lines[0]]['total'] += 1
else:
#print lines
if len(lines[2]) > 0: # with year data
edge.append(lines[2].strip())
hist.add(lines[2].strip())
if not stat.has_key(lines[0]):
stat.update({lines[0] : {'total' : 1, 'wyear' : 1}})
else:
stat[lines[0]]['total'] += 1
stat[lines[0]]['wyear'] += 1
else: # without year data
if not stat.has_key(lines[0]):
stat.update({lines[0] : {'total' : 1, 'wyear' : 0}})
else:
stat[lines[0]]['total'] += 1
edge_list_all.append(edge)
else:
print "invalid line!", lines
f.close()
# # statistical
# f = open("../result/result_may28/me/statistics/year_statistical.csv","w")
# f.write("univ,total,wyear\n")
# for key in stat:
# f.write("%s,%d,%d\n" %(key, stat[key]['total'], stat[key]['wyear']))
# f.close()
# index, dist, cdf = hist.cdf()
# print hist._max, hist._min
# print len(index), index
# print len(dist), dist
# print len(cdf), cdf
# print sum(dist)
#
# # the CDF of year distribution
# f = open("../result/result_may28/ee/statistics/year_cdf.csv","w")
# f.write("year,freq,percentile\n")
# for i in range(len(index)):
# f.write("%s,%d,%.3f\n" %(index[i], int(dist[i]), cdf[i]))
# f.close()
#
# exit(0)
# univlist = sorted(s.iteritems(), key = lambda asd:asd[0], reverse = False)
# fo = open("../data/out_me.csv","w")
# for i in univlist:
# fo.write("%s,%d\n" %(i[0],i[1]))
# fo.close()
# exit(0)
## re-organize the edge with weights
edge_dict = {}
for edge in edge_list_all:
key = edge[0]+"#"+edge[1]
if edge_dict.has_key(key):
edge_dict[key] += 1.0
else:
edge_dict.update({key : 1.0})
edge_list = []
for item in edge_dict.iteritems():
if self_edge == True:
edge = []
edge.extend(item[0].split("#"))
edge.append(item[1])
edge_list.append(edge)
else:
edge = []
nodes = item[0].split("#")
if not nodes[0] == nodes[1]:
edge.extend(nodes)
edge.append(item[1])
edge_list.append(edge)
node_list = sorted(s.keys(), reverse = False)
return node_list, edge_list
if child.tag.endswith('when'):
ts = dateutil.parser.parse(child.text)
when = int(time.mktime(ts.timetuple()))
if prev_when is not None:
dts.append(when - prev_when)
prev_when = when
rmv = RollingMeanVar(0.001)
Y = []
for i, dt in enumerate(dts):
rmv.update(dt, i)
Y.append(rmv.mean())
pylab.plot(range(len(Y)), Y, color=RED)
setupPlot(
'Sample Index',
'Interval (s)',
'Polling Interval - Rolling Average'
)
pylab.savefig('timings-frequency.png')
pylab.close()
Y = map(operator.itemgetter(1), histogram(dt % 60 for dt in dts))
pylab.bar(range(len(Y)), Y, width=1.0)
setupPlot(
'Interval (s, mod 60)',
'Frequency',
'Polling Interval - Seconds mod 60'
)
pylab.savefig('timings-second-histogram.png')
pylab.close()
def keyword_distribution(self, rm_en=True):
f = open(
"results/statistics_hashtag_sentiscores_%d.csv" %
len(self.keywords), "w")
f.write(
"keyword,user_count,tag_count,count,sw_avg_neg,sw_avg_pos,sw_count_neg,sw_count_zero,sw_count_pos,sw_min,sw_max,"
+
"st_avg_neg,st_avg_pos,st_count_neg,st_count_zero,st_count_pos,st_min,st_max,"
+
"sc_avg_neg,sc_avg_pos,sc_count_neg,sc_count_zero,sc_count_pos,sc_min,sc_max\n"
)
for word in self.keywords:
hsw = stats.histogram()
hst = stats.histogram()
has = stats.histogram()
for score in self.keywords[word]["scores"]:
hsw.add(score["sentiword_score"])
hst.add(score["sentiment_score"])
has.add(score["sentiscore"])
f.write("%s,%d,%d,%d,%.3f,%.3f,%d,%d,%d,%.3f,%.3f," %
(word, self.keywords[word]["user_count"],
self.keywords[word]["tag_count"], hsw._count,
hsw._mean_neg, hsw._mean_pos, hsw._count_neg, hsw._zero,
hsw._count_pos, hsw._min, hsw._max))
f.write("%.3f,%.3f,%d,%d,%d,%.3f,%.3f," %
(hst._mean_neg, hst._mean_pos, hst._count_neg, hst._zero,
hst._count_pos, hst._min, hst._max))
f.write("%.3f,%.3f,%d,%d,%d,%.3f,%.3f\n" %
(has._mean_neg, has._mean_pos, has._count_neg, has._zero,
has._count_pos, has._min, has._max))
distribution_sw = hsw.histogram_2()
distribution_st = hst.histogram_2()
distribution_as = has.histogram_2()
# sentiword score
x_axis = range(len(distribution_sw))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i * 0.05 for i in range(42)]
plt.bar(x_axis,
distribution_sw,
width=0.8,
facecolor='blue',
alpha=0.5)
X_ticks = ["%.1f" % (x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i - 1) % 2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Sentiword Score")
plt.ylabel("Percentage (%)")
plt.title("Sentiword Score Distribution - %s" % word)
plt.savefig(OUT_DIR + SUB_DIR_1 + "%s.png" % word, dpi=50)
plt.clf()
# sentiment score
x_axis = range(len(distribution_st))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i * 0.05 for i in range(42)]
plt.bar(x_axis,
distribution_st,
width=0.8,
facecolor='blue',
alpha=0.5)
X_ticks = ["%.1f" % (x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i - 1) % 2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Sentiment Score")
plt.ylabel("Percentage (%)")
plt.title("Sentiment Score Distribution - %s" % word)
plt.savefig(OUT_DIR + SUB_DIR_2 + "%s.png" % word, dpi=50)
plt.clf()
# sentiscore score
x_axis = range(len(distribution_as))
rcParams['figure.figsize'] = 24, 5
x_ = [-1.025 + i * 0.05 for i in range(42)]
plt.bar(x_axis,
distribution_as,
width=0.8,
facecolor='blue',
alpha=0.5)
X_ticks = ["%.1f" % (x_[i]) for i in range(len(x_))]
for i in range(len(X_ticks)):
if (i - 1) % 2 != 0:
X_ticks[i] = ""
plt.xticks(x_axis, X_ticks)
plt.xlim(0, len(x_axis))
plt.grid(True)
plt.xlabel("Avg Senti Score")
plt.ylabel("Percentage (%)")
plt.title("Senti Score Distribution - %s" % word)
plt.savefig(OUT_DIR + SUB_DIR_3 + "%s.png" % word, dpi=50)
plt.clf()
print "done %s: neutral_st %d neutral_sw %d neutral_as %d" % (
word, hst.count_zero(), hsw.count_zero(), has.count_zero())
f.close()