Making great tasting dog food with Python
How we are using social network analysis to understand palatability in dog food recipes
Sunday 16th, 11:30 (Room A)
I’m a data scientist working for a tailor made dog food company, and in this session I’ll explain how we are applying ideas in social network analysis to understand and improve the overall taste and palatability of our dry dog food recipes.
Understanding what makes dog food delicious is a difficult problem to solve, and as the intended consumers of our food are dogs, this presents some obvious limitations to the methods available to approach this problem.
I am taking a fairly unique and data led approach to this, using ideas in social network analysis to visualise our dog food recipes and find ingredients (or combinations of ingredients) common in recipes where we know that the dog didn’t like the recipe.
In this talk, I’ll tell the story of how this project evolved, and how we’ve been using the Python library NetworkX to help us create some really insightful visualisations of our data. Along the way I’ll cover some of the great things about this library, as well as some of the pitfalls.
- The speaker suggested this session is suitable for data scientists.