Automation of the Updated Food Label Information Program (FLIP 2020): A Comprehensive Canadian Branded Grocery and Restaurant Food Composition Database
Bibliographic record
Abstract
Traditional methods for creating food composition databases struggle to cope with the large number of products and the rapid pace of turnover in the food supply. The objective is to overview the updated Food Label Information Program (FLIP2020), a big data approach for the evaluation of the Canadian food supply and present the latest methods used in the development of this database. The University of Toronto's Food Label Information Program (FLIP) is a database of Canadian prepackaged and chain restaurant foods and beverages collected since 2010. FLIP 2020 was developed using website “scraping” and machine learning (ML) coupled with artificial intelligence-enhanced optical character recognition (AI-OCR) to collect and manage food labelling information (e.g., nutritional composition, price, product images, ingredients, brand, etc.) on all foods and beverages available on seven major Canadian e-grocery retailer websites and 201 Canadian chain restaurants between May 2020 and February 2021. FLIP 2020 is comprised of 74,445 prepackaged food products and 21,225 menu items available on websites of seven retailers, 2 location-specific duplicate retailers and 141 chain restaurants. Food products were classified under multiple national and international categorization systems, in order to analyse similar foods under different systems. Of 57,006 food and beverage products available on seven retailers’ websites, nutritional composition data were available for about 60% of the products and ingredients were available for about 45%. Data for energy, protein, carbohydrate, fat, sugar, sodium and saturated fat were present for 54–65% of the products, while fibre information was available for 37%. Of the 201 eligible chain restaurants with ≥ 20 national outlets, 70% provided nutritional information. All provided energy, 84% provided saturated fat, total sugar and sodium, and 50% provided all 13 required nutrients listed on the Nutrition Facts table. FLIP, with its comprehensive sampling and granularity and use of ML/AI-OCR, is a powerful tool for evaluating and monitoring the Canadian food supply environment. This research was supported by funds from a Canadian Institutes of Health Research Project Grant.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.008 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.009 | 0.009 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.004 | 0.003 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.010 | 0.007 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".