Mapping the data: what, why, how?

Jul 20, 2020 | Data Governance

As data is massively collected by the company, it must be exploited – that is, analyzed, sorted, segmented, and then qualified. But because it is not confined to a fixed and immutable base, like a trophy in a showcase, this data must necessarily be made accessible and used by everyone within the structure, without technical compartmentalization or limitations.


What is data mapping?

In essence, data mapping is a process for identifying and visualizing data entry and processing points. This form of topography is common to all of the company’s information systems, which gives all employees the opportunity to use it – and to speak the same “data” language, both among Data Scientists and IT managers and within business departments, who are less expert in the field.

This map is drawn using three main “utensils”:

The semantic tool, which consists of listing in a business glossary the metadata of data and business objects specific to the company, in order to optimize the understanding and context of the data.
Data models, providing precise indications on how data is modeled and stored in different storage systems (structured, semi-structured or unstructured). Another tool is consubstantial to it: the technical data dictionary.
The data flow processing design tool, providing information on the methods of data transformation, manipulation and processing through the company’s various information systems.
To these three tools must be added the formalization of the format for making data available, access and conditions of use, so that the sharing of mapping does not come up against obstacles relating to confidentiality and integrity of data understanding.


Why is this process crucial?

Gaining a 360° view of the mass of data in circulation in a company: this is the solution offered by the data mapping process. The goal is to universalize it (at least on the scale of the structure), i.e. to make it totally accessible and understandable to all employees, so that everyone is able to identify the origin of a data, measure its calculation method, identify possible redundancies, etc.

Imagine for a moment: you want to find a customer’s billing address. Not only can the sources of this information be multiple, but its very definition can differ according to the services from which it comes. The very understanding of “address” can vary greatly depending on company glossaries, and the inability to track a history of updates will contribute to your being lost in an inextricable network of databases – with no assurance of being able to get your hands on what you’re looking for.

Mapping data therefore allows you to create links between the technical vision (the storage and transformation of data in IT applications) and the business vision (the use made of it within different departments). This process aims to break down historical silos, to decompartmentalize services – in short, to build a common space in which all employees speak the same language, in order to give everyone the tools for better operational data governance.

How to get the most out of data mapping?

A good data mapping requires a pedagogical process: it is a question of transforming a historically siloed structure (technical partitions separate each department, limiting the easy circulation of information and uses within the company) into a living, evolving collective knowledge, modelled by the employees themselves, according to their experience. 

This process cannot be limited to a basically top-down approach, which would condemn all mapping to obsolescence as soon as it is launched. It can only take place in an iterative way, with one project and initiative after another, at the heart of a global approach of integration into daily tasks. For example, everyone can participate at his or her own level in the mapping of business granular data; and only then will it be possible to create bridges between personal initiatives, in order to broaden the scope and produce aggregated data, based on common characteristics. A resolutely bottom-up approach

Ultimately, the ambition is to obtain a cartography encompassing all the services and applications of the IS, offering a 360° vision of the data circulating in the structure. And allowing everyone to make the most of it.