I’m an intern at LABEXIA, an agri-Food laboratory located in Quimper (29). This organization is one of production sites of GROUPE CARSO and it employs over 150 people.


Analyzes are carried out at the request of customers, whether it’s for human food, animal feed, vegetation or cosmetic products.



Data management and Data rendering

Labexia processes more than 1 million samples per year, on which are performed, on average, 3 million tests which yield 5 million results.

My project is focused on this BIG DATA perspective. With the IT team, I aim to find a new BI (Business Intelligence) tool in order to make available, for key users (laboratory managers, business segments and customer service), dashboards and pivot tables ad hoc.

Ad hoc: concerned or dealing with a specific subject, purpose, or end.

A BI tool helps to extract, to qualify, to model and to exploit data from many data sources (servers, ERP, Excel…) It has provided an opportunity to edit reporting, dashboards, monitor laboratory activity or even to spot extreme and erroneous data without calling on computer services for punctual data extraction by SQL queries.


BI process


Process: major milestones

Before any implementation of new computer programs, it’s necessary to collect, exhaustively, the end-user’s needs.

Following this step, we drafted specifications.

Then we have selected and tested different Business Intelligence solutions (Qlik, Power BI, ReportOne, Tableau) to establish benchmarking based on key criterion (ETL* incorporated directly, use of variables, web portals, bookmark selections…)

ETL is short for extract, transform, and load, three database functions that are combined into one tool to pull data out of multiple databases (SQL server, Excel, CSV…) and place into another database with changes like filters, date format, alias ,joins.

It’s preferable to design models based on relatively different products, to offer users a range of solutions. One example is QlikView and Qlik Sense, the former requires computer skills and is very complete to create dashboards. However, QlikSense is user-focused with the ‘’drag-and-drop’’ function but it has fewer functionalities like warnings on critical thresholds.

During these stages, it’s essential to contact software companies and providers for help and advice, and for future potential collaborators.

This project will enable users to know the steps for the establishment of new software and more specifically, a BI TOOL.

Andy A.







Laisser un commentaire