Analysis of the phenomenon of absenteeism in the production center.

The Cerealto Siro Group aims to develop a first use case of advanced analytics, based on the analysis of absenteeism in its production centers, as the first concrete application of its Big Data strategy that will serve as a driver of a global strategy in this area.

To achieve this, a robust and scalable architecture will be defined to drive the Group's digital transformation and, more specifically, decision making based on data analysis.

Data integrating the project and architecture:

Data:                                                                                  

Job characterization data    

Data on employees' transfer transactions       

Contextual data by work center                     

Absenteeism historical data

Architecture:

Virtual machine in Azure cloud environment (Microsoft)

Postgresql DB

Power BI Visualization (Microsoft)

HOW WE APPROACH IT

The determination of the variables that influence absenteeism is a process that has matured in different studies of classical industrial sociology. The aim of advanced analytics is to establish the main determining variables and the optimal way to reduce absenteeism with the minimum possible impact on operational processes and the work environment.

Methodologically, we initiated an RFM analysis that provides us with an initial classification of employees into different groups according to 3 variables:

• Recency: the number of days that have elapsed since a worker has been absent from the workplace.
• Frequency: the number of days between one absence due to absence and the next.
• Amount: in this case we will apply two considerations. On the one hand, the number of days of accumulated absenteeism in the employee's history and, on the other, the business cost of said absenteeism.

The analysis of the Group's historical absence data allows us to distribute the entire employee base into a classification of 27 groups, according to the 3 variables analyzed, scoring employees with a value from 1 to 3 according to their frequency of absence, recency and amount.

 As a complement to the RFM analysis, we will evaluate the weight of the different variables that characterize each employee in the practice of absenteeism. This diagnostic analysis assigns a weight to these variables so that the data and the algorithms applied to the data can detect patterns of common characteristics among the employees who practice absenteeism.

Prior to the development of the use case itself and according to Smartup's methodology in the development of projects in the field of Big Data, it will be necessary to undertake a series of common tasks as well as the definition and development of a scalable infrastructure that allows modeling and analysis for this first use case and beyond:

• Inventory of data sources
• Representation Data model
• Data dictionary
• Exploratory Analysis
• Data quality audit

WHAT WE DO

Meet

• IOB (Internet Of Behavior)
• Big Data Analytics 
• RFM
• Power BI
• Dashboard

Connect

Convert

Capitalize

• Automation

WHAT RESULTS WE OBTAINED

After the project, the Cerealto Siro Group has the capacity to develop strategies for the reduction of absenteeism from multiple internal actions, for which a tool for internal and external use is enabled through which it can transfer knowledge about this specific behavior over time. 

Globally, with this initiative, the company is carrying out the necessary tasks to coordinate internally the path towards a data-driven company with which to manage the evolution of its production teams.