Research Area
Computational Gastronomy
Computational Gastronomy is the data science of food — applying network analysis, machine learning, and flavor chemistry to understand how we cook, eat, and nourish ourselves. It is a young field, largely unknown outside a handful of research labs, and dramatically undervalued in the gastronomic sciences.
I studied Gastronomic Sciences at UNISG Pollenzo, where the relationship between food, culture, and science is taken seriously. That education shaped a conviction: the knowledge embedded in artisanal food traditions deserves the same computational rigor we give to genomics or climate science. We share this work to advance the field — and to invite institutions, researchers, and innovators to build it with us.
Open Call #001
Accepting collaboratorsStructural Patterns in Italian Starred Cuisine: A Network Analysis of Ingredient Pairing, Menu Composition, and Culinary Identity Across Michelin-Starred Restaurants
FTB Open Research
Italy holds the most Michelin-starred restaurants in Europe, yet the structural patterns underlying their menus remain unexplored through a computational lens. This research applies network science, flavor compound analysis, and statistical methods to investigate whether Italian starred restaurants share common ingredient pairing signatures, menu architectures, or compositional principles that distinguish them from non-starred establishments.
Building on the flavor network framework (Ahn et al., 2011), the recipe similarity network methodology (Bellingeri et al., 2025), and the culinary grammar formalism (Bagler, 2022), this study aims to:
- Construct ingredient co-occurrence networks from starred restaurant menus across Italian regions and identify structurally central dishes and ingredients
- Test whether starred Italian cuisine exhibits distinctive food pairing patterns compared to traditional and non-starred repertoires
- Map the tension between regional culinary identity and creative innovation — how far do starred chefs deviate from their regional culinary grammar?
- Identify “bridge ingredients” that connect traditional Italian flavor profiles to contemporary fine-dining techniques
The dataset will comprise publicly available tasting menus, carte, and seasonal menus from one- to three-star restaurants across a minimum of five Italian regions. Analysis will combine network centrality metrics, community detection, and statistical comparison with reference networks built from traditional Italian cookbooks.
Who we are looking for
- Food science & gastronomic studies departments
- Network science & data analysis researchers
- Gastronomic universities (UNISG, Basque Culinary Center, etc.)
- Independent researchers & PhD candidates
- Institutions interested in funding food data science
Follow this research
Get updates on datasets, methodology, and results as they develop.
Future Directions
Open research questions
Directions we believe are worth pursuing. If you are working on any of these — or something adjacent — reach out.
Flavor Networks for Italian Cuisine
Map the anti-pairing pattern of regional Italian cuisines and quantify structural uniqueness.
Artisanal Quality as Signal
An inverse FPro — a craft score distinguishing artisanal products chemically from industrial.
Italian Cookbook Networks
Network science on regional cookbooks to identify characterizing recipes and ingredient DNA.
Reading Library
State of the art
Annotated reading notes on the research shaping computational gastronomy — from the foundational papers that launched the field to the latest work in rice metabolomics and network analysis of cuisines.
These are our reading notes and analysis. All original works belong to their respective authors and publishers — each entry links to its original source.
Foundational
3 papersThe papers that defined the field
Computational Gastronomy: Capturing Culinary Creativity by Making Food Computable
Ganesh Bagler & Mansi Goel
The definitive field overview of Computational Gastronomy — from flavor networks to AI recipe generation to sustainability — introducing the Turing Test for Chefs with an F1 score of 69.88%.
A Generative Grammar of Cooking
Ganesh Bagler
The first formal generative grammar for cooking, treating recipes as combinatorial systems of culinary concepts (cubits) assembled by syntactic rules — making the creative act of cooking computable.
Flavor Network and the Principles of Food Pairing
Yong-Yeol Ahn, Sebastian E. Ahnert, James P. Bagrow, Albert-László Barabási
The seminal paper that launched computational gastronomy — introducing the flavor network (381 ingredients, 1,021 compounds) and revealing that Western cuisines pair flavor-sharing ingredients while East Asian cuisines avoid them.
Network Analysis
1 paperGraph theory applied to culinary systems
The Recipe Similarity Network
Michele Bellingeri, Axel Bidon-Chanal Badia, Marta Vila Rigat, Roberto Alfieri, Massimiliano Turchetto, Davide Cassi
Network science meets Catalan cuisine — a normalized similarity measure and clique-based community detection reveal the structural DNA of traditional and haute cuisine cookbooks, identifying Allioli and Joan Roca's signature Becada as characterizing recipes.
Nutrition & Health
1 paperMolecular networks connecting diet and disease
Decoding the Foodome: Molecular Networks Connecting Diet and Health
Giulia Menichetti, Albert-László Barabási, Joseph Loscalzo
Reveals the 'dark matter of nutrition' — food contains 139,000+ chemicals beyond the ~150 tracked by national databases. Introduces FPro, a machine learning score that quantifies food processing from nutrient patterns.
Rice Science
6 papersAroma, metabolomics, and quality characterization
Quality Characteristics of Rice Varieties and Their Suitability for Bibimbap
Hyeonbin Kim, Eun Ah Sim, Chang-Min Lee et al.
A template for computational dish optimization — matching rice variety characteristics to bibimbap requirements through integrated physicochemical analysis, texture profiling, and consumer preference, revealing that moderate amylose (17-18%) and texture stability are key.
Comparative Study of Key Aroma Components in Rice of Different Aroma Types
Shengmin Qi, Haibin Ren, Haiqing Yang, Lianhui Zhang, Min Zhang
Moves beyond the aromatic/non-aromatic binary — classifying 11 rice varieties into three aroma types (sweet, cereal, complex) using flavor metabolomics, with 8 differential volatile compounds and GC-IMS fingerprinting for rapid identification.
Analysis of Rice Characteristic Volatiles and Their Influence on Rice Aroma
Shuimei Li, Hongyan Li, Lin Lu et al.
Identifies the three key volatiles governing rice aroma across 31 varieties: 2-AP (positive, popcorn), hexanal (negative, aging marker), and trans,trans-2,4-decadienal (negative) — establishing a dual-marker quality framework.
Unique Metabolic Profiles of Korean Rice
Yujin Kang, Bo Mi Lee, Eun Mi Lee et al.
Metabolomics deconvolutes the factors shaping rice chemistry: polishing explains 37.5% of variance, variety 13.5%, and cultivation region 4.6% — the first study to simultaneously quantify all three, with 127 of 156 metabolites enriched in brown rice.
Science and Technology of Aroma, Flavor, and Fragrance in Rice
Deepak Kumar Verma, Prem Prakash Srivastav (Editors)
The definitive reference book on rice aroma science — 11 chapters covering extraction methods (SDE, SPME, SCFE), genetic engineering of fragrance, aroma biomarkers, and the BADH2 gene underlying 2-AP biosynthesis.
Chemical, Physical, Textural and Sensory Evaluation on Italian Rice Varieties
C. Simonelli, L. Galassi, M. Cormegna, P. Bianchi
Italy's national rice authority characterizes ten varieties (Carnaroli, Arborio, Vialone Nano...) through chemistry, texture, and sensory profiling — the first systematic sensory evaluation of Italian rice, revealing three groups by amylose content.