19 June 2025
AI Applied to Sales Forecasting!
We are proud to announce that ALIT Technologies has collaborated with a group of students on a project involving time-series analysis applied to the sales of a chemical product. This initiative created a valuable connection between industry and academia, offering students the chance to apply predictive models to real-world data, and giving our company the opportunity to explore innovative analytical approaches. The study leveraged Artificial Intelligence to enhance accuracy and refine predictive capabilities.
Data by Riccardo Vendraminetto, Alex Meggiolaro and Paolo Fabris
An Analytical and Forecasting Project
Riccardo Vendraminetto, Alex Meggiolaro, and Paolo Fabris, students from the “Data Science” course at the University of Padua, conducted an analysis of the monthly sales of an ALIT chemical product from 2018 to 2024. The goal of their research was to identify trends, seasonal effects, and patterns through advanced statistical models such as linear regression, ARIMA, and exponential smoothing methods.
Their exploratory analysis revealed a clear upward trend in sales, with irregular seasonality and increasing variability in recent years. In particular, May and September showed the highest average sales, while April and August recorded significantly lower levels. These findings required careful modeling to uncover underlying market dynamics.
Among the tested models—supported by AI—the most effective linear regression model was the one accounting for both general sales trends and seasonal variations, achieving good predictive accuracy (adjusted R² of 0.64). However, due to inconsistent fluctuations in the data, mathematical corrections were necessary to achieve more reliable results.
At the same time, ARIMA models, which analyze historical data to predict future trends, demonstrated excellent forecasting capability. Ultimately, the best-performing model was a combination of linear regression and ARIMA, significantly improving predictive accuracy, reducing errors, and making forecasts more reliable.
Data by Riccardo Vendraminetto, Alex Meggiolaro and Paolo Fabris
A Concrete Example of Industry-Academia Collaboration
This project, enhanced by Artificial Intelligence, illustrates how collaboration between businesses and academic students can produce meaningful outcomes. ALIT Technologies provided real-world data, enabling students to face the complexities of an industrial setting. In return, students could apply advanced mathematical methodologies to an actual scenario, measuring effectiveness through metrics like RMSE and MAE.
An interesting aspect highlighted by the study was the comparison between traditional forecasting models and innovation diffusion methodologies, such as the Bass model and the Guseo-Guidolin Model (GGM). While the Bass model suggested a slowdown in future sales growth, the GGM offered a more optimistic view, highlighting the importance of imitation in product adoption processes.
A First Step Toward New Opportunities
The collaboration between ALIT Technologies and the students has demonstrated the substantial value generated when academic research meets industry needs. Besides offering a high-level learning experience for the students, this project has opened new horizons for ALIT in terms of sales data analysis and demand forecasting.
This partnership could pave the way for future joint initiatives, encouraging ongoing dialogue between academia and industry to effectively address market challenges. Integrating new explanatory variables, applying machine learning techniques, and exploring emerging patterns in data represent just a few possible directions for this fruitful collaboration.
Artificial Intelligence: A Strategic Asset for the Future
The use of Artificial Intelligence in this project, clearly demonstrates its value beyond manufacturing processes. AI has become essential for accurate sales forecasting, thanks to its capability to rapidly process large amounts of data. AI-driven forecasting models can continuously adapt to market fluctuations, changes in raw material costs, and shifts in demand. Leveraging this innovation properly can deliver a significant strategic advantage, empowering companies to make informed and agile business decisions.
