Impact
Define all Pernod Ricard business data glossary, list the source and their availability. Be able to retrieve per project the availability and freshness of data to run algorithms.
Context
In an AI transformation, the data knowledge and literacy is key: what is the right name and definition of the data to collaborate and find the right data for the specific business case, where is the data available in different market and database.
Usually, a project team is doing all this documentation for its own project, and keep it locally in Excel sheet. When a new project start, the same work is done again and again. The goal of Magellan, was to consolidate 3 elements: The data dictionary, the Governance information, and the data sources to accelerate new AI solutions creation.
User research showed that project team are willing to documentate those information if the tool provide them some benefits for their own project. For this purpose we created a new « Project view » which enable users to list all the data they need and see in real time all data already available to create the algorithm.
We wanted to go on step further for the tool to be helpful daily and not only in the build phase. For that, Magellan display for each source of data its freshness. Consequence is that when project leaders need to run algorithm and get fresh results they know easily if the data is fresh or no.
More information about the digital transformation in Pernod Ricard FY21 Annual Report Page 34
Final screens
Homepage – Catalog browsing using iTunes inspirations
The user problem when using data is that they don’t always knows the data classification. The product should adapt to their needs and let them find the right data whatever they remind: a data domain, family, sub-family, or field name. Here the navigation behave like a multi-filter selection and user can start from any category.
Data field page: glossary, governance and sources
Project pages
Indicate the fields needed for a specific project, get the sources available. Project leader of each solution are informed of the freshness of the data if the final users needs to use the AI solution to take new decisions.
Navigation
The user research shown that most often collaborators don’t know the exact name of the data and even less the classification defined by the HQ. To ease the search, in addition to the global search bar, we wanted to enable users to reach data whatever classification or name they remember. For this purpose we used the same filters navigation than Apple use in iTunes,. But with a difference: our filters navigation works both ways!