def daily_correlation(self):
"""
Determines the pmcc between days since account creation and usage + displays linegraph with line of best fit.
"""
delta = datetime.datetime(2020, 4, 27) - datetime.datetime(
2017, 7, 19) + datetime.timedelta(days=1)
daily_data = self._data_time(slice(0, 1),
"",
slice(0, 10),
return_ints=False)
days_since_account_creation = list(range(delta.days))
daily_activity = []
for i in range(delta.days):
date = datetime.datetime.strftime(
datetime.datetime(2017, 7, 19) + datetime.timedelta(days=i),
"%Y-%m-%d")
#print(date, i, daily_data.count(date))
daily_activity.append(daily_data.count(date))
print(
self.ylabel,
round(
np.corrcoef(days_since_account_creation, daily_activity)[0][1],
4))
util.graph((days_since_account_creation, daily_activity),
"all time",
"Day",
ylabel=self.ylabel,
unique=True,
line_of_best_fit=True)
util.graph_histogram_2(daily_activity, ylabel=self.ylabel)
def months(self, year):
"""
Conducts analysis on months.
Parameters
----------
year : string
The data to match - can be YYYY or empty for all time.
"""
time_data = self._data_time(slice(0, len(year)), year, slice(5, 7))
if len(year) == 4:
util.graph(time_data,
year,
"Month",
xtick_min=1,
xtick_max=13,
ylabel=self.ylabel)
else:
year = "all time"
util.graph(time_data,
year,
"Month",
xtick_min=1,
xtick_max=13,
ylabel=self.ylabel)
#util.graph_boxplot(time_data,year,"Month")
util.table(time_data,
print_latex=self.print_latex,
print_lists=self.print_lists)
def graph_polarity_rating(l): x_label = 'Polarity Score' y_label = 'Rating (stars)' title = 'Polarity of Review Text vs. Reviewer Rating' save_filename = 'output/polarity_vs_stars.png' graph(l, title, x_label, y_label, save_filename) return
def graph_strlen_rating(l): x_label = 'String Length' y_label = 'Rating (stars)' title = 'Length of Review Text vs. Reviewer Rating' save_filename = 'output/review_length_vs_rating.png' graph(l, title, x_label, y_label, save_filename) return
def graph_day_polarity(l):
x_label = "Day of Week"
y_label = "Polarity Score"
title = "Day of Week vs. Polarity Score"
save_filename = "output/day_of_week_vs_polarity_score.png"
graph(l, title, x_label, y_label, save_filename)
return
def graph_day_rating(l):
x_label = "Day of Week"
y_label = "Rating (stars)"
title = "Day of Week vs. Rating (stars)"
save_filename = "output/day_of_week_vs_rating.png"
graph(l, title, x_label, y_label, save_filename)
return
def days_range(self, start_date, finish_date):
"""
Conducts analysis on all days between start_date and finish_date inclusive.
Parameters
----------
start_date : string
The date to match - YYYY-MM-DD.
finish_date : string
The date to match - YYYY-MM-DD.
"""
time_data = []
for year in range(int(start_date[:4]), int(finish_date[:4]) + 1):
if year == int(start_date[:4]):
start_month = int(start_date[5:7])
else:
start_month = 1
if year == int(finish_date[:4]):
finish_month = int(finish_date[5:7])
else:
finish_month = 12
for month in range(start_month, finish_month + 1):
year_month = str(
year) + "-" + "0" * (2 - len(str(month))) + str(month)
if year_month == start_date[0:7]:
start_day = int(start_date[8:10])
else:
start_day = 1
if year_month == finish_date[0:7]:
finish_day = int(finish_date[8:10])
else:
finish_day = 31
for day in range(start_day, finish_day + 1):
date = year_month + "-" + "0" * (2 -
len(str(day))) + str(day)
time_data += self._data_time(slice(0, 10),
date,
slice(0, 10),
return_ints=False)
util.graph(time_data,
start_date + " to " + finish_date,
"Day",
ylabel=self.ylabel)
#util.graph_boxplot(time_data,start_date+" to "+finish_date,"Day")
print("Between", start_date, "and", finish_date, "inclusive:")
util.table(time_data,
print_latex=self.print_latex,
print_lists=self.print_lists)
def years(self):
"""
Conducts analysis on years.
"""
time_data = self._data_time(slice(10, 11), "T", slice(0, 4))
util.graph(time_data,
"all time",
"Year",
xtick_min=self.min_year,
xtick_max=self.max_year + 1,
ylabel=self.ylabel)
util.table(time_data,
print_latex=self.print_latex,
print_lists=self.print_lists)
def days(self, year_month):
"""
Conducts analysis on days.
Parameters
----------
year_month : string
The data to match - YYYY-MM.
"""
time_data = self._data_time(slice(0, 7), year_month, slice(8, 10))
util.graph(time_data, year_month, "Day", ylabel=self.ylabel)
util.graph_boxplot(time_data, year_month, "Day")
util.table(time_data,
print_latex=self.print_latex,
print_lists=self.print_lists)
def breaks(self, date, min_break=datetime.timedelta(days=1)):
"""
Conducts analysis on breaks I have taken from Instagram.
Parameters
----------
date : string
When should the analysis cover - empty string for all time.
min_break : datetime.timedelta, optional
What is the minimum break that should be included. The default is datetime.timedelta(days=1).
"""
time_data = self._data_time(slice(0, len(date)),
date,
slice(0, 19),
return_ints=False)
time_data.sort(reverse=True)
time_data = [
datetime.datetime.strptime(datestamp, "%Y-%m-%dT%H:%M:%S")
for datestamp in time_data
]
max_break = datetime.timedelta(seconds=0)
break_start = ""
breaks = ([], [])
for i in range(len(time_data) - 1):
break_length = time_data[i] - time_data[i + 1]
break_length_hours = break_length.days * 24 + break_length.seconds // 3600
if break_length > min_break:
breaks[0].append(time_data[i].strftime("%Y-%m-%d"))
breaks[1].append(break_length_hours)
breaks[0].reverse()
breaks[1].reverse()
util.graph(breaks,
util.date_to_time_period(date),
"day",
ylabel="hours break",
unique=True)
util.table(breaks,
sort_by_likes=True,
unique=True,
print_latex=self.print_latex,
print_lists=self.print_lists)
