IRDb Presentation

by the SWEdish Chefs

Transcript

1. Internet Recipe Database (IRDb) Handpicked, mixed, and sautéed by the SWEdish Chefs
2. Agenda • Team (SWEdish Chefs) • Project (IRDb) • Demonstration • SWEet Music Critique • IRDb Critique
3. SWEdish Chefs • Role: Front-end • Commits: 244 • Issues opened: 34 • Unit tests: 2 Min Bi Chef de cuisine • Roles: Database, unit tests, API, documentation • Commits: 66 • Issues opened: 20 • Unit tests: 35 Joshua Cannon Sous chef • Roles: Back-end, Docker/Carina deployment, some front-end, documentation • Commits: 72 • Issues opened: 24 • Unit tests: 0 Marshall Gallatin Chef de partie • Roles: Database, API design, documentation • Commits: 71 • Issues opened: 20 • Unit tests: 0 K.C. Hawes-Domingue Legumier • Roles: API design, documentation, SWEet Music page • Commits: 16 • Issues opened: 4 • Unit tests: 0 Victor Yap Rôtisseur
4. Our project: IRDb • Built on two fundamental models: Recipes Ingredients
5. What can you do with IRDb? • Look up ingredients in a recipe • See what recipes use a particular ingredient • Find recipes that meet certain dietary restrictions • View nutritional information on ingredients • … and more!
6. Tools
7. Data Sources
8. RESTipe API • /recipes • page • limit • cuisine • diet • /recipes/{id} • /recipes/ingredientSearch • ingredients • /ingredients • page • limit • /ingredients/{id} • /ingredients/{id}/nutrition • /ingredients/{id}/recipes • page • limit
9. Demonstration: http://swedishchef.me
10. SWEet Music Critique http://sweetmusic.me
11. SWEet Music What did they do well? • Visually appealing • Clean layout • Nice color palette and fonts • Seamless Spotify integration • Provided Artist News under Artist Pages
12. SWEet Music What did we learn from their website? • You don’t need a lot of fancy stock photos and animations to make a website look good • Having a load page (“Searching…”) is great for UX – informs user that the website is still working when it needs extra time to perform a search
13. SWEet Music What can they do better? • Some of the Wiki is difficult to read and was inconsistent with the rest of the Wiki • Allow searching by track length in the API • Preferably using a time range (e.g., “2:00-3:00”)
14. SWEet Music What puzzles us about their website? • Does the drop-down menu on the front page’s search bar work? (“Search artists”, “Search albums”… ) • Wiki is unclear about whether the data found on the website is static, or if their scraper.py script still runs every so often
15. IRDb Critique
16. Internet Recipe Database What did we do well? • Our data model is sound and the navigation of our website reflects that • Recipes ↔ Ingredients • We have lots of data that you probably won’t find aggregated in the same way anywhere else
17. Internet Recipe Database What did we learn? • SQLAlchemy was great for combining Python with SQL queries while only needing to know Python • Content created by the public is difficult to clean up for use in a database • Stop worrying so much about GMOs • Ex.: Strawberries are a hybrid species created in the 1750s
18. Internet Recipe Database What can we do better? • Pick data sources that work better together • Massaging data took too long due to natural language variation • Integrate nutritional data (calorie count, etc.) into each recipe page • Polish the overall look and feel
19. Internet Recipe Database What puzzles us? • Sorting all columns of data across pages • Why does the USDA nutritional database have information for ostrich leg meat, ground squirrel, and walrus liver, but not common grocery store items such as chili powder, garlic salt, and Nutella?
20. Happy cooking and bon appétit!