"""
This module contains helper functions for the EDA and modelling
exercises. Feel free to use them to get you started more quickly.
"""
import pandas as pd
import numpy as np
from datetime import datetime
def parse_time(s):
return datetime.strptime(s, "%Y-%m-%d").date()
def read_demand(path):
df = pd.read_csv(path)
df = df.assign(date=lambda df: df.date.apply(parse_time))
df = df.set_index("date")
df.index = pd.DatetimeIndex(df.index)
return df
def read_promotions(path):
df = pd.read_csv(path, index_col=0)
df = df.assign(promotion_date=lambda df:
df.promotion_date.apply(parse_time))
df = df.set_index("promotion_date")
df.index = pd.DatetimeIndex(df.index)
return df
def clean(ts: pd.Series) -> pd.Series:
# Replaces missing values
return ts.bfill().fillna(ts.mean())
def clean_demand_per_group(demand: pd.DataFrame) -> pd.DataFrame:
"""TODO add docstring"""
sus = demand.supermarket.unique()
skus = demand.sku.unique()
for su in sus:
for sku in skus:
demand.loc[(demand.sku == sku) & (demand.supermarket == su),
"demand"] = clean(demand.loc[(demand.sku == sku) & (demand.supermarket ==
su), "demand"])
return demand
def merge(demand: pd.DataFrame, promotions: pd.DataFrame) ->
pd.DataFrame:
promotions = promotions.rename_axis("date").assign(promotion=True)
demand = demand.merge(
promotions,
on=["supermarket", "sku", "date"],
how="outer",
)
demand = demand.assign(promotion=lambda df:
df.promotion.fillna(False))
return demand
def extend_promotions_days(promotions, n_days):
""" Extends the promotions to have multiple rows for a specific
number of days.
The input promotions is assumed be specified with a single row with a
starting date.
The output extends the input promotions with multiple days, one row
for each day of the promotion.
"""
n_promotions = len(promotions)
initial_promotions = promotions.copy()
promotion_id = np.arange(n_promotions)
extended_promotions =
promotions.copy().assign(promotion_id=promotion_id)
for days_to_add in range(1, n_days):
additional_promotion_days =
initial_promotions.copy().assign(promotion_id=promotion_id)
additional_promotion_days.index += pd.Timedelta(days_to_add, "d")
extended_promotions =
extended_promotions.append(additional_promotion_days)
return extended_promotions
def aggregate_to_weekly(df):
grouped = df.groupby(["sku", "supermarket"])
# Performs a simplistic aggregation of promotion. If a promotion
occured during the week this variable will be true.
weekly = grouped.apply(lambda df: df.resample("W").agg({"demand":
"sum", "promotion": "max"}))
weekly = weekly.reset_index().set_index("date")
return weekly
