Machine Learning for Retail Sales Forecasting — Features Engineering

Understand the impacts of additional features related to stock-out, store closing date or cannibalization on a Machine Learning model for sales forecasting

Features Engineering for Machine Learning for Retail Sales Forecasting — (Image by Author)

Discover the power of Machine Learning for retail sales forecasting with features engineering.

In this article, we explore how additional features like stock-out, store closing dates, and cannibalization impact the accuracy of a Machine Learning model using the M5 Forecasting competition dataset.

Learn how features engineering can improve your sales forecasting by up to 60%, as shown in the Makridakis Forecasting Competitions.

Follow our experiment and analysis of six different feature buckets to optimize your sales predictions and implement effective inventory management rules.

Introduction

Based on the feedback of the last Makridakis Forecasting Competitions, Machine Learning models can reduce the forecasting error by 20% to 60% compared to benchmark statistical models. (M5 Competition)

Their major advantage is the capacity to include external features that heavily impact the variability of your sales.

For example, e-commerce cosmetics sales are driven by special events (promotions) and how you advertise a reference on the website (first page, second page, …).

This process called features engineering is based on analytical concepts and business insights to understand what could drive your sales.

In this article, we will try to understand the impact of several features on the accuracy of a model using the M5 Forecasting competition dataset.

💌 New articles straight in your inbox for free: Newsletter
📘 Your complete guide for Supply Chain Analytics: Analytics Cheat Sheet

SUMMARY
I. Introduction
1. Data set
2. Initial Solution using LGBM
3. Features Analysis
II. Experiment
1. Additional features
2. Results
III. Conclusion and next steps

I. The M5 Forecasting Dataset: Overview and Objectives

1. Data set of Retail Sales Transactions

This analysis will be based on the M5 Forecasting dataset of Walmart store's sales records (Link).

• 1,913 days for the training set and 28 days for the evaluation set
• 10 stores in 3 states (USA)
• 3,049 unique in 10 stores
• 3 main categories and 7 departments (sub-category)

The objective is to predict sales for all products in each store, in the following 28 days right after the available dataset. We have to perform 30,490 forecasts for each day in the prediction horizon.

M5 Forecasting Competition Dataset — (Image by Author)

We’ll use the validation set to measure the performance of our model.

2. Initial Solution using Machine Learning Algorithm LGBM

As a base model, we will use a very clear and concise notebook shared by Anshul Sharma in Kaggle. (Link)

The idea is to understand how we can improve the accuracy of the model only by adding additional features (without touching the hyperparameters or changing the algorithm).

In this notebook, you will find all the different steps to build a quite good model with a reasonable computing time:

1. Import and processing of raw data
2. Exploratory Data Analysis
3. Features Engineering
   i) Seasonality: week number, day, month, day of the week
   ii) Pricing: the weekly price of an item in each store, special events
   iii) Trends: sales lags (n-p days), average volume per {item, (item +store)}, …
   iv) Categorical Variables encoding: item, store, department, category, state
4. Model Training: 1 model LightGBM per store

3. Features Engineering to improve the model

In order to emphasize the impact of features engineering, we will not change the model and only look at which features we use.

Let us split the features used in this notebook into different buckets.

—
Bucket 1: Transactional Data

# Item id
'id', 'item_id', 
# Store, Category, Department
'dept_id', 'cat_id', 'store_id', 'state_id'
# Transaction time
'd', 'wm_yr_wk', 'weekday', 'wday', 'month', 'year'
# Sales Qty, price and promotional events
'sold', 'event_name_1', 'event_type_1', 'event_name_2', 'event_type_2', 'sell_price'

events and sell_price
Capture the impact on sales of a special event on an item of selling price XXX.

What could be the impact of a special event with -20% reduction on sales of baby formula the second week of the month?

Open Question
What would be the impact on the accuracy if we do one-hot encoding for the categorical features?

—
Bucket 2: Sales Lags and Average

# Sales lag n = sales quantity of day - n
'sold_lag_1', 'sold_lag_2', 'sold_lag_3', 'sold_lag_7', 'sold_lag_14', 'sold_lag_28'
# Sales average by 
'item_sold_avg', 'state_sold_avg',
'store_sold_avg', 'cat_sold_avg', 'dept_sold_avg',
# Sales by XXX and YYYY
'cat_dept_sold_avg', 'store_item_sold_avg', 'cat_item_sold_avg',
'dept_item_sold_avg', 'state_store_sold_avg',
'state_store_cat_sold_avg', 'store_cat_dept_sold_avg'

lags
Measure the week-on-week or month-on-month (7 days, 28 days) similarities to capture the periodicity of sales due to people shopping at these frequencies.

Do you have relatives going to the hypermarket every Saturday to shop for the whole week?

💡 Follow me on Medium for more articles related to 🏭 Supply Chain Analytics, 🌳 Sustainability and 🕜 Productivity.

II. Experiment: Features Engineering Strategies for an Improved Accuracy

Find the full code in my Github repository: Link (Follow me :D)
My portfolio with other projects: Samir Saci

1. Additional features

Based on business insights or common sense, we will add additional features, built with existing ones, to help our model to capture all the key factors impacting your customer demand.

Experiment to understand the impact of features on the error with the validation set — (Image by Author)

—
Bucket 3: Rolling Mean and Rolling Mean applied on lag

# Rolling mean on actual sales
'rolling_sold_mean', 'rolling_sold_mean_3', 'rolling_sold_mean_7',
'rolling_sold_mean_14', 'rolling_sold_mean_21', 'rolling_sold_mean_28'
# Rolling mean on lag sales
'rolling_lag_7_win_7', 'rolling_lag_7_win_28', 'rolling_lag_28_win_7', 'rolling_lag_28_win_28'

rolling_sold_mean_n
Measure the average sales of the last n days.

Rolling mean is sometimes used alone as a benchmark model for statistical forecasting.

Code

rolling_lag_n_win_p
Measure the average sales of a p days windows ending n days ago.

Code

BUSINESS INSIGHTS
Sunglasses seasonality

If the rolling mean of the last 7 days is 35% higher than the average sales of the week before that means you have started the summer season.

—
Bucket 4: Sales Trend and Rolling Maximum

# Selling Trend
'selling_trend', 'item_selling_trend', 
# Rolling max
'rolling_sold_max', 'rolling_sold_max_1', 'rolling_sold_max_2', 'rolling_sold_max_7', 'rolling_sold_max_14', 'rolling_sold_max_21',
'rolling_sold_max_28'

Selling trend
Measure the gap between the daily sales and the average.

Code

Rolling max
What is the maximum sales in the last the n days?

Code

Spoiler: this feature will have an important impact on your accuracy.

—
Bucket 5: Stock-Out and Store Closed

# Stock-out id
'stock_out_id'
# Store closed
'store_closed'

stock-out

Explain that you have zero sales because of stock availability issues.

Code

—

Bucket 6: Price Relative with the same item in other stores or other items in the sub-category

# Relative delta price with the same item in other stores
'delta_price_all_rel'
# Relative delta price with the previous week
'delta_price_weekn-1'
# Relative delta price with the other items of the sub-category
 'delta_price_cat_rel'

delta_price_weekn-1
Capture the price evolution week by week.

BUSINESS INSIGHTS
Promotions for Slow Movers

In order to reduce their inventory and purge slow movers, stores may apply aggressive pricing to boost sales.

delta_price_all_rel: Sales Cannibalization at store level
Several stores competing for sales of the same item because of price difference.

delta_price_cat_rel: Sales Cannibalization at sub-category level
Several items of the same sub-category competing for sales.

Code

2. Results

After running a loop of training with the six different buckets (using the same hyperparameter with the Kaggle notebook) we have the following results:

RMSE on the validation set for each of the step of experiment— (Image by Author)

—

STEP 1 to STEP 2: -29% RMSE Error

Features Importance — (Image by Author)

Sales lags are positively impacting the accuracy of your model:.

BUSINESS INSIGHTS
Your sales of today is highly impacted by the sales of the previous days.

—

STEP 2 to STEP 3: -118% RMSE Error

Features Importance — (Image by Author)

BUSINESS INSIGHTS
The top 3 features are all related to the sales of the last three days.

Question
Based on this insight, what could be the performance of a model like Exponential Smoothing who is taking a ponderate sum of the previous sales to compute the forecasts.

—

STEP 3 to STEP 4: -12% RMSE Error

Features Importance — (Image by Author)

Rolling max features are taking the lead at the top of the features.

—

STEP 4 to STEP 5: -0.1% RMSE Error

Features Importance — (Image by Author)

BAM!
I am devastated to see that the potential main added value of this article, showing the impact of stock-out or store closing, has a limited impact on the accuracy of the model.

—

STEP 5 to STEP 6: -1.75% RMSE Error

Features Importance — (Image by Author)

The model accuracy is slightly better but we do not see any of the added features in the top 20.

Samir SACI - Data Science for Supply Chain Portfolio

III. Conclusion and next steps

In conclusion, this article highlights the importance of features engineering in improving the accuracy of Machine Learning models for retail sales forecasting.

By including external features such as stock-out and store closing, businesses can better understand the factors impacting their customer demand.

Through this analysis, we have shown the positive impact of sales lags, rolling max, and other features on the accuracy of the model.

1. Understand the results

The results in terms of model accuracy are quite satisfying.

However, there is still some frustration with not having found any correlation between some of the newly added features and the model performance.

Therefore, the next step will be to work on these features and the model (let’s remember that we did not touch the initial model here) to check if there is any possibility of using these features to better forecast your sales.

If these business insights are not improving your forecasts we need to understand why.

2. Implement Inventory Management Rules

Now that you have your forecasting model, you need to implement an inventory management rule to manage the replenishment of your stores.

You can take inspiration from these three articles where we try to implement rules assuming a deterministic or stochastic demand.

Inventory Management for Retail — Stochastic Demand

Inventory Management for Retail — Deterministic Demand

Inventory Management for Retail — Periodic Review Policy

About Me

Let’s connect on Linkedin and Twitter, I am a Supply Chain Engineer that is using data analytics to improve logistics operations and reduce costs.

If you are interested in Data Analytics and Supply Chain, have a look at my website

Samir Saci | Data Science & Productivity

References

• Time Series Forecasting-EDA, FE & Modelling📈, Anshul Sharma, Link
• “Machine Learning for Retail Demand Forecasting”, Towards Data Science, Samir Saci
• “Inventory Management for Retail — Stochastic Demand”, Towards Data Science, Samir Saci
• “Inventory Management for Retail — Deterministic Demand”, Towards Data Science, Samir Saci
• “4 Impacting Projects to Start Your Data Science for Supply Chain Journey”, Towards Data Science, Samir Saci