def mean_cust_per_shop_if_holiday(df, data_from=None):
if data_from is None:
data_from = df
df['MeanCustPerShopIfHoliday'] = p.Series(np.zeros(len(df)), df.index)
means = dict()
num = dict()
for i in data_from.index.tolist():
prom = d.content_of(data_from, 'IsHoliday', i)
reg = d.content_of(data_from, 'StoreID', i)
index = str(reg)
val = d.content_of(data_from, 'NumberOfCustomers', i)
if prom != 0:
try:
means[index] += val
num[index] += 1
except KeyError:
means[index] = val
num[index] = 1
for i in df.index.tolist():
prom = d.content_of(df, 'IsHoliday', i)
reg = d.content_of(df, 'StoreID', i)
if prom != 0:
index = str(reg)
df.set_value(i, 'MeanCustPerShopIfHoliday', means[index]/num[index])
return df
def mean_std_sales_per_shop_per_day(df, data_from=None):
if data_from is None:
data_from = df
df['MeanSalesPerShopPerDay'] = p.Series(np.zeros(len(df)), df.index)
df['StdSalesPerShopPerDay'] = p.Series(np.zeros(len(df)), df.index)
means = dict()
stds = dict()
num = dict()
for i in data_from.index.tolist():
id = d.content_of(data_from, 'StoreID', i)
day = d.content_of(data_from, 'Day', i)
index = str(id) + day
val = d.content_of(data_from, 'NumberOfSales', i)
try:
means[index] += val
num[index] += 1
except KeyError:
means[index] = val
num[index] = 1
for i in df.index.tolist():
id = d.content_of(df, 'StoreID', i)
day = d.content_of(df, 'Day', i)
index = str(id) + day
df.set_value(i, 'MeanSalesPerShopPerDay', means[index]/num[index])
for i in data_from.index.tolist():
id = d.content_of(data_from, 'StoreID', i)
day = d.content_of(data_from, 'Day', i)
index = str(id) + day
val = d.content_of(data_from, 'NumberOfSales', i)
try:
stds[index] += (val - means[index]/num[index]) * (val - means[index]/num[index])
except KeyError:
stds[index] = (val - means[index]/num[index]) * (val - means[index]/num[index])
for i in df.index.tolist():
id = d.content_of(df, 'StoreID', i)
day = d.content_of(df, 'Day', i)
index = str(id) + day
df.set_value(i, 'StdSalesPerShopPerDay', np.sqrt(stds[index]/num[index]))
return df
def one_hot_numeric(ds, attr, header):
"""Transforms the given attribute of the given DataFrame object into one hot encoding.
If you plan to use this, don't use split attribute.
Returns a DataFrame object."""
vals = d.values_of(ds, attr)
new_cols = vals
for new in new_cols:
ds[header + str(new)] = p.Series(np.zeros(len(ds)), ds.index)
for i in ds.index.tolist():
if d.content_of(ds, attr, i) == new:
ds.set_value(i, header + str(new), 1)
return ds
def mean_sales_per_month_per_region(df, data_from=None):
if data_from is None:
data_from = df
df['MeanSalesPerRegionPerMonth'] = p.Series(np.zeros(len(df)), df.index)
means = dict()
num = dict()
for i in data_from.index.tolist():
month = int(d.content_of(data_from, 'Month', i))
reg = d.content_of(data_from, 'Region', i)
index = str(month) + "_" + str(reg)
val = d.content_of(data_from, 'NumberOfSales', i)
try:
means[index] += val
num[index] += 1
except KeyError:
means[index] = val
num[index] = 1
for i in df.index.tolist():
month = int(d.content_of(df, 'Month', i))
reg = d.content_of(df, 'Region', i)
index = str(month) + "_" + str(reg)
df.set_value(i, 'MeanSalesPerRegionPerMonth', means[index]/num[index])
return df
def mean_cust_per_month_per_shop(df, data_from=None):
if data_from is None:
data_from = df
df['MeanCustPerShopPerMonth'] = p.Series(np.zeros(len(df)), df.index)
means = dict()
num = dict()
for i in data_from.index.tolist():
month = int(d.content_of(data_from, 'Month', i))
reg = d.content_of(data_from, 'StoreID', i)
index = str(month) + "_" + str(reg)
val = d.content_of(data_from, 'NumberOfCustomers', i)
try:
means[index] += val
num[index] += 1
except KeyError:
means[index] = val
num[index] = 1
for i in df.index.tolist():
month = int(d.content_of(df, 'Month', i))
reg = d.content_of(df, 'StoreID', i)
index = str(month) + "_" + str(reg)
df.set_value(i, 'MeanCustPerShopPerMonth', means[index]/num[index])
return df
def add_min_per_shop(df, data_from=None):
if data_from is None:
data_from = df
ids = d.values_of(data_from, 'StoreID')
stds = dict()
for id in ids:
try:
_ = stds[str(id)]
except KeyError:
stds[str(id)] = min_per_shop(data_from, id)
df['min_shop'] = p.Series(np.zeros(len(df)), df.index)
for i in df.index.tolist():
df.set_value(i, 'min_shop', stds[str(d.content_of(df, 'StoreID', i))])
return df
def add_avg_cust_per_shop(df, data_from=None):
if data_from is None:
data_from = df
ids = d.values_of(data_from, 'StoreID')
means = dict()
for id in ids:
try:
_ = means[str(id)]
except KeyError:
means[str(id)] = average_cust_per_shop(data_from, id)
df['meancustshop'] = p.Series(np.zeros(len(df)), df.index)
for i in df.index.tolist():
df.set_value(i, 'meancustshop', means[str(d.content_of(df, 'StoreID', i))])
return df
def one_hot(ds, attr, header, split=False):
"""Transforms the given attribute of the given DataFrame object into one hot encoding.
If you plan to use this, don't use split attribute.
Returns a DataFrame object."""
vals = d.values_of(ds, attr)
if split:
new_cols = []
for v in vals:
split = v.split("-")
for s in split:
if not new_cols.__contains__(s):
new_cols.append(s)
else:
new_cols = vals
for new in new_cols:
ds[header + new] = p.Series(np.zeros(len(ds)), ds.index)
for i in ds.index.tolist():
if d.content_of(ds, attr, i).find(new) != -1:
ds.set_value(i, header + new, 1)
return ds
