Avisia - Using Dataiku to Predict the Outcome of a Beauty Contest

Ophe_T

Name: Ophélie Tourlan

Title: Data Analyst Consultant

Country: France

Organization: Avisia

Founded in 2007, over the years AVISIA has become a reference player in data, digital, and technologies. Through our team of enthusiasts, we are specialized in consulting, integration, and the realization of data-centric projects. Our fundamental values include:

• The expertise that allows us the recognition of the market. To this end, we regularly train our employees so that they acquire new skills.
• The benevolence that materializes for our employees through great proximity and listening to their needs and desires, but also for our customers to whom we guarantee efficiency and pragmatism.
• Disruption through a daily desire to build innovative data projects.

Awards Categories:

• Best Moonshot Use Case

Business Challenge:

The goal was to build a model to predict the beauty pageant "Miss France."

Business Solution:

We created a Dataiku project with different data sources: Insee, Wikipedia, and Twitter. After using different techniques like web scraping and topic modeling, we created a model which led us to a prediction score to determine which Miss France contestant had the highest probability of winning the contest.

Dataiku enabled us to collaborate as a team on the project, organize our sources through zones, and easily use both code and visual recipes. Scenarios enabled us to automatically run the model to get a new prediction daily. Updating the Flow each year is made easier by using variables.

Business Area: IT/Cybersecurity/Data

Use Case Stage: Built & Functional

Value Generated:

This project helped us to gain visibility thanks to media interest. Many articles were written about us and our project during the month of the contest, which included coverage on Vanity Fair, TF1 Info, Voici, Public, Cosmopolitan, and Grazia. We were also invited by BFM Business to present our project and talk about our company.

Value Brought by Dataiku:

Dataiku enables us to work efficiently as a team and to develop our skills through lessons at the Dataiku Academy. Scenarios enable us to automatically run the model to get a new prediction every day, which saves us precious time. Updating the Flow each year is also made easier by using variables.

Value Type:

• Increase revenue
• Save time
• Increase trust