def plot_day_of_week_all(data, tar_months):
for tar_month in tar_months:
print('Month: %s' % dh.month_no_to_name(tar_month))
for data_name in data.keys():
df = data[data_name]
print('dataset:%s, Monday to Sunday in %s' %
(data_name, dh.month_no_to_name(tar_month)))
plot_day_of_week(df, tar_month, data_name)
plt.show()
def plot_stats_weekly(data):
n_data = len(data)
weekday_names = [dh.week_no_to_name(i) for i in range(7)]
n_cols = 2
n_rows = int(np.ceil(n_data / float(n_cols)))
for tar_month in [1, 5, 7, 10]:
print('month %d' % tar_month)
fig = plt.figure(tar_month, figsize=(6 * n_cols, 4 * n_rows))
for i, (data_name, data_df) in enumerate(data.items(), 1):
df_month = dh.get_df_month(data_df, tar_month)
mins = []
maxs = []
means = []
stds = []
meds = []
for tar_day in range(7):
df_weekday = dh.get_df_dayofweek(df_month, tar_day)
mins.append(df_weekday.min().values[0])
maxs.append(df_weekday.max().values[0])
means.append(df_weekday.mean().values[0])
stds.append(df_weekday.std().values[0])
meds.append(df_weekday.median().values[0])
ax = fig.add_subplot(n_rows, n_cols, i)
plot_stats(mins, maxs, means, stds, meds, ax, '-')
month_name = dh.month_no_to_name(tar_month)
_ = ax.set_title('dataset:%s, month:%s' % (data_name, month_name))
_ = ax.set_xticks(range(7))
_ = ax.set_xticklabels(weekday_names, rotation=90)
_ = ax.set_ylim(ymin=0)
plt.tight_layout()
plt.show()
def test_no_to_str(self):
self.assertEqual(dh.week_no_to_name(-1), 'Unknown')
self.assertEqual(dh.week_no_to_name(0), 'Mon')
self.assertEqual(dh.week_no_to_name(1), 'Tue')
self.assertEqual(dh.week_no_to_name(2), 'Wed')
self.assertEqual(dh.week_no_to_name(3), 'Thu')
self.assertEqual(dh.week_no_to_name(4), 'Fri')
self.assertEqual(dh.week_no_to_name(5), 'Sat')
self.assertEqual(dh.week_no_to_name(6), 'Sun')
self.assertEqual(dh.week_no_to_name(7), 'Unknown')
self.assertEqual(dh.month_no_to_name(-1), 'Unknown')
self.assertEqual(dh.month_no_to_name(1), 'January')
self.assertEqual(dh.month_no_to_name(4), 'April')
self.assertEqual(dh.month_no_to_name(7), 'July')
self.assertEqual(dh.month_no_to_name(10), 'October')
self.assertEqual(dh.month_no_to_name(13), 'Unknown')
def plot_seasonaly(data, tar_months):
for data_name, data_df in data.items():
plot_src = {}
print(
'dataset:%s, daily consumptions in %s' %
(data_name, ', '.join([dh.month_no_to_name(m)
for m in tar_months])))
for i, m in enumerate(tar_months, 1):
df_month = dh.get_df_month(data_df, m)
plot_src.update({m: dh.reshape_day_by_day(df_month)})
ymin = np.nanmin([x.values.min() for x in plot_src.values()])
ymax = np.nanmax([x.values.max() for x in plot_src.values()])
n_rows = 2
n_cols = len(tar_months)
fig = plt.figure(figsize=(len(tar_months) * 6, n_rows * 4))
for i, (m, v) in enumerate(sorted(plot_src.items()), 1):
ax = fig.add_subplot(n_rows, len(tar_months), i)
plot_daily_all(ax, v, ymin, ymax, i, m, data_name)
ax = fig.add_subplot(n_rows, len(tar_months), i + len(tar_months))
plot_daily_ave_med_std(ax, v, ymin, ymax, m, data_name)
# ax = fig.add_subplot(n_rows, n_cols, i + 2 * len(tar_months))
# plot_daily_ave(ax, v, ymin, ymax, m, data_name)
plt.show()
def plot_stats_data_comparison(data):
min_list = [data[k].min().values[0] for k in data.keys()]
max_list = [data[k].max().values[0] for k in data.keys()]
mean_list = [data[k].mean().values for k in data.keys()]
std_list = [data[k].std().values[0] for k in data.keys()]
med_list = [data[k].median().values[0] for k in data.keys()]
n_cols = 2
n_rows = 3
fig = plt.figure(1, figsize=(6 * n_cols, 4 * n_rows))
ax = fig.add_subplot(n_rows, n_cols, 1)
plot_stats(min_list, max_list, mean_list, std_list, med_list, ax)
_ = ax.set_xticks(range(len(data)))
_ = ax.set_xticklabels(data.keys(), rotation=90)
_ = ax.set_ylim(0, 65000)
_ = ax.set_title('min, max, mean and std of all months')
for i, tar_month in enumerate([1, 4, 7, 10], 3):
min_list = []
max_list = []
mean_list = []
std_list = []
med_list = []
for k in data.keys():
dfm = dh.get_df_month(data[k], tar_month)
min_list.append(dfm.min().values[0])
max_list.append(dfm.max().values[0])
mean_list.append(dfm.mean().values[0])
std_list.append(dfm.std().values[0])
med_list.append(dfm.median().values[0])
ax = fig.add_subplot(n_rows, n_cols, i)
plot_stats(min_list, max_list, mean_list, std_list, med_list, ax)
month_name = dh.month_no_to_name(tar_month)
_ = ax.set_title('month:%s' % month_name)
_ = ax.set_xticks(range(len(data)))
_ = ax.set_xticklabels(data.keys(), rotation=90)
_ = ax.set_ylim(0, 65000)
plt.tight_layout()
def plot_diff_days_hours_each_month(data, tar_months):
for data_name, data_df in data.items():
for month_no in tar_months:
df_month = dh.get_df_month(data_df, month_no)
dname_month = '%s(%s)' % (data_name, dh.month_no_to_name(month_no))
plot_diff_days_hours(dname_month, df_month)