Business ethics at the heart of our values

Ellisphere is a company with a significant data asset. Over the course of 125 years, we have accumulated a plethora of data. We regularly upgrade our processing as technology offers us new capabilities. Ethics and responsibility are central to our data processes.

These changes are guided either by the continuous evolution of new technologies, or in a more disruptive way by rethinking in particular certain modes of sourcing or refining information.

As soon as we implement a new solution for our customers, involving new data, we have to ask ourselves questions about its life cycle and the management of its obsolescence data by data. The rules of historization and purging are taken into account from the start of our developments.

 

A resolutely ethical data processing organization

Data for Ellisphere can be considered as the essential raw material in the construction of our services. Our Data Management entity is structured in several technical poles. Each of them interacts with our business experts, thus mixing technical, functional and mathematical skills to :

 

Sourcing

In building our solutions, we have always favored quality over quantity when it comes to data. How can we enrich our document collections with useful, complementary and consistent data? How can we increase our relevance to the core target audience of the most consulted companies?

Each implemented data must answer these questions in order to guarantee its relevance within our database and thus bring an enrichment to the existing.

Faced with the multiplicity of formats and the rapid development of Open Data, we are increasingly confronted with quality problems. Therefore, we have introduced new graphical tools for data exploration, qualification and rectification into our processing chains. The objective? To think about new working methods between our business experts and our IT experts to make the development and industrialization of our future data processing and exploration processes more fluid.

 

Prepare

Over the last few years, we have completely redeveloped our data preparation process through different phases:

  • Standardization: indispensable notion considering the heterogeneity of raw data.
  • Consolidation: to check the consistency of the data between them.
  • Referencing: to feed coherent and ready-to-use broadcast views.

This exercise allowed us to identify and classify all of our documents in a complete and referenced dictionary.

 

Innovate and provide tools

A dedicated center of expertise has the mission of validating, through POC (Proof of Concept), the interest of introducing new processing capacities thanks to the technology. This stage is designed to respond to various business problems. This unit is at the heart of our latest innovations with the deployment in production of sub-systems based on big data and event-driven streaming. In addition, other projects are currently being studied, such as the use of graph technology.

 

Ethical data, a driver for innovation

Since 2016, we have invested heavily in R&D by equipping ourselves with a new architecture that allows us to set up a Fab Lab in data science and Artificial Intelligence (AI). The implementation of AI in our solutions has allowed us to innovate in our markets.

Indeed, since the release of our 3rd generation score, we have greatly accelerated our ability to produce customized scores, decision models, and indicators based on self-learning algorithms. The explicability of these models is at the heart of our concerns and remains central to the business ethics to which we respond.

For each algorithm produced by data science, we guarantee the explicability of each decision made by the algorithms in conjunction with our business experts. Explanatory phases, financial ratios, sectoral data... All these elements will help to understand and explain the results provided by our algorithms.

We have also introduced agility into our AI chain, with an iterative Learning/Industrialization/Observation process. The data science team calibrates the algorithms in a lab disconnected from production and industrializes the code produced directly on our validation platforms, then on our MLOps production platforms (Machine Learning and Operations).

From one end of the design chain to the other, Ellisphere's mission is to be a responsible player in ethical data. Rooted in our history, we strive to respect a strong business ethic in the collection, storage, processing and transformation of data.