Data Science in the bakery
A small bakery is usually not associated with a place where the latest technology in the field of process prediction is being introduced. Data Science is the latest trend in management and optimization.
The development of computing and access to endless information has opened the door to extraordinary paths of development, and today the biggest barriers to development are our mentality, our thinking horizons.
The wave of data science is over!
It's a data center for Data Science, a lot of data, it's an environment filled with information about manufacturing and sales processes, customer and supplier behavior, social phenomena, and logistics processes.
In a previous article on Data Science, I showed you how the digital revolution has taken over the world of the peripheral bakery.
One of the most exciting areas of data science is what's called Machine Learning, and the most famous and spectacular achievement in this field is Tesla's self-driving cars.
Machine learning has been applied to countless systems in the field of diagnostics and process maintenance, creating something in the form of continually learning artificial intelligence, and machine learning is also widely used in prediction.
These people (called Data Scientists) create complex models that autonomously and continuously learn, adapting to an infinite flow of data.
The data science revolution is not a material development, because at the beginning of this century, computers already had enough computing power for this kind of activity, and their further development is not a necessary condition for the development of data science.
W-MOSZCZYNSKI-2019-2-02Ordering from the bakery
Once, almost 10 years ago, I was asked to help with a difficult logistical task.
These were elaborate locations. Suburbs and towns used to come by train to work. Every morning a great river of people would roam around the bakeries to come back early in the evening.
The bakery shops were well supplied before dawn, the problem was with the second, the evening delivery. The early bakery supply was almost completely bought out. Unfortunately, there were no rules about what to bring to the second shift. Sometimes the bakery and the cookies were missing, other times most of the products had to be baked at night.
The inefficiencies of the second change had a significant impact on the profitability of the bakery and the candy store.
Wise Observations
The bakery's managers noted that there were certain rules for evening customers, such as the two basic types of bread best sold on Mondays and Wednesdays, and the most chocolate was sold on Fridays.
It was also observed that bad weather was favorable for buying cookies and coffee, and the worst day of sales was Thursday, with shops experiencing a sales boom before long weekends, and it was also observed that each day the wave of returns began at a slightly different time, which affected the delivery times to the shops.
This information was used in the completion of supplies and thus significantly reduced the returns of the products of the second amendment.
Sir, head's not enough.
But there was a dilemma: how much to send on Thursday, the weakest day of demand, when the long weekend is coming, the period of high demand, and each store had a dozen products delivered on the second shift, and each store had a slightly different set of customer behavior rules.
And behaviours have also changed over time, they have evolved in a way, and the factors that influence customer behaviour have varied, and how much to bring to Bronewski when it's bad weather and Thursday, but at the same time, it's a big name day?
Dozens of stores, with dozens of products for seven days, produce an artisanal quantity of about 2,000 decisions taking into account different conditions, each with its own conditions.
Often, at the same time, there are conflicting conditions, as one of the staff at the completion said: 'Sir, the catch is not enough.
With the help of econometrics, World War II was a period of dynamic development of econometric methods.
It was intended to provide the American forces fighting on both fronts, with limited transport and storage resources, with what they need and as much as they need, on time, while minimizing losses and threats.
It was during this time that linear regression modeling and logistics models began to be used on a massive scale, and these tools are now being used by Data Scientist in Machine Learning.
The Excel sheet that cured the bakery
Back at our bakery, the task wasn't too complicated. It was to collect all the sales transactions from the last few years, from all the stores and taking into account the dates, times and indexes of the products sold.
The data was then added directly from the Internet to the data – temperature, pressure, day of the week and several other factors that could potentially affect sales.
Even the oldest Excel comes with a set of tools that can quickly separate data sets into stores and product indexes into descriptive factors, and every Excel has a linear regression model generator.
The hardest part was designing an Excel spreadsheet, which, based on the daily data, updated dozens of regression models, and each store had its own Excel spreadsheet that had a list of products, and all you had to do was write the delivery date, and the sheet said how many items to send to the store.
A prerequisite for the proper functioning of the program was the daily replenishment of a database on which regression models were based.
Data science is not a field that's far from life, but it's a field that allows you to work on boring, endless calculations where you always lose to a machine.
Wojciech Moszczyński — graduate of the Department of Econometrics and Statistics of Nicolaus Copernicus University in Toruń; specialist in econometrics, finance, data science, and management accounting. He specializes in the optimization of production and logistics processes. He conducts research in the area of the development and application of artificial intelligence. For years he has been engaged in the popularization of machine learning and data science in business environments.
Dodaj komentarz