Forecasting models: 2-Wk Proof of Concept

DESCRIPTION: Design an appropriate and scalable solution allowing you to obtain reliable business previsions (financial, demand, supply chain, quality, and other forecasting scenarios) by melting the most advanced analytical methodologies with the Azure AI-Machine Learning and data visualization ecosystem as ML Studio and Notebook. Within a well-designed time series analysis, the proposed AI forecasting PoC will be structured in order to perform a preliminary Exploratory Data Analysis (EDA) on a selected sample of data and summarize the main characteristics (e.g. pattern detection, spot outliers, hypothesis test) through an analytical notebook. The effective description of the underlying fundamental factors determining the observed cases represents an important added value of the solution.

OUTCOMES:

• Contextual understanding of the selected Azure analytical tools
• Discover how ML and AI could be exploited to improve the considered process
• Model explanation to extract meaningful information from the data

DELIVERABLES:

• Setup of Azure analytical solutions to operate ML analysis
• Overview of Azure ML and AI capabilities in conjunction with the data visualization features
• Data model build and training
• AML notebook about exploratory data analysis and forecasting model main results
• Next steps plan to extend the POC capabilities to other businesses scenarios

PLANNING: DAY 1 - Briefing about the POC scope and high-level details of setting up the Azure analytical ecosystem. Data selection and understanding session with Business team DAYS 2–9 - Set up of AI and ML environment. Onboard datasets and configure model parameters. Train and test the forecasting models and run predictions. Develop a model explainability notebook highlighting the most important features (where applicable) DAY 10 - Roadmap based on POC results and insights

AUDIENCE: ICT Dept., Data Scientists, Business Analysts

LANGUAGES: Italian or English