Rice Dishes of the World Kathryn Haydon and Andy Mauromoustakos University of Arkansas Data from Wikipedia 265 rice recipes gt70 countries worldwide represented Asia 1 in unique dishes More than 90 of the worlds rice is produced and consumed in Asia ID: 1045349
Download Presentation The PPT/PDF document "Case Study of Text Explorer" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
1. Case Study of Text ExplorerRice Dishes of the WorldKathryn Haydon and Andy MauromoustakosUniversity of Arkansas
2. Data from Wikipedia: 265 rice recipes
3. >70 countries worldwide represented
4. Asia: #1 in unique dishesMore than 90% of the world’s rice is produced and consumed in Asia
5. Rice puddings and non-pudding dishesMost of the dishes here are not rice puddingsBut most rice puddings are not from Asian countries, so they are unique as rice dishes overallThere are enough puddings to justify separate analysis
6. Rice puddings are concentrated in European, Middle Eastern, South and Central American countries, and IndiaCan puddings be predicted by description?
7. Text explorer of all dishes
8. 2-cluster analysis of all dishesTop Terms by Cluster12TermScoreTerm Scorefri·2.57milk8.51chicken2.42sugar·7.62meat·2.03cinnamon6.01veget·2.01coconut4.09onion·1.45raisins2.91fish1.45vanilla2.07bean·1.33butter1.79tomato·1.29dessert1.69spice·1.12sweet1.44pork1.05sweetened1.32Mosaic PlotVery few puddings are grouped into Cluster 1Most puddings are in Cluster 2, but so are quite a few non-puddingsThere are non-pudding rice desserts in Cluster 2Separation of Puddings by Clustering
9. Split text exploration by puddingNon-PuddingRice Pudding
10. 3-cluster analysis of rice puddingsTop Terms by ClusterMosaic Plot123TermScoreTerm ScoreTerm Scorecardamom1.51sugar12.92raisins14.23black1.51cinnamon5.23vanilla12.32saffron1.51milk4.03milk10.82slow-boiled1.51brown·2.37soaked7.83flavored1.01sweet2.37rum7.77anise1.01flour2.37butter7.71rosewater0.90fruit·2.24coconut7.59cold0.89lemon1.84peel5.89dessert0.88sauce1.78nutmeg5.85nuts0.76salt·1.78cloves5.83Cluster 1 includes flavors typical of Indian and Middle Eastern cooking: cardamom, saffron, anise, rosewaterCluster 2 emphasizes sweetness, cinnamon, and fruitCluster 3 contains warm flavors like vanilla, nutmeg, and cloves, as well as rum and coconut that suggest Caribbean flavors
11. Rice pudding cluster analysis, mapped:Cluster 1 is associated with the Middle East, Southern Asia, and Eastern European countriesCluster 2 is associated with Western Europe and some Eastern European countriesCluster 3 is dominated by countries in the Western Hemisphere
12. Non-pudding rice dish analysisMore dishes, more complicatedClear dominance of Asian cuisine
13. 4-cluster analysis of non-puddingCluster CharacteristicsTop Terms by Cluster1234TermScoreTerm ScoreTerm ScoreTerm Scorecake·0.70coconut6.49paste5.15sauce8.67soup0.68milk4.48tomato·4.70chicken7.74spice·0.67glutinous2.90sweet4.62onion·6.27egg·0.57wrap·2.75pepper·3.89veget·5.95bread·0.56sticky2.74red3.88meat·5.61curri·0.53steam·2.65white3.87fish5.22lentil·0.52boil·2.13bean·3.54season·4.54flour0.49sugar·2.12oil3.51salt·4.43breakfast0.30shred·1.98onion·3.09garlic3.74pea·0.16leaf1.88savory2.34slice·3.67Cluster 1 populated by flour-containing foods (cakes and breads), and basics like soup, eggs, lentilsCluster 2 includes sticky rice dishesCluster 3 is savory and spicy, but vegetarian Cluster 4 is dominated by proteins like chicken and fish and flavors like onion and garlic
14. Asian countries occupy >50% of every clusterMosaic PlotPatterns in the DataA simple clustering scheme does not separate by region neatlyNon-Asian countries most represented in Cluster 1Many basic food terms in Cluster 1Great diversity among Asian countries and cuisinesJust looking at Asian countries might show some patterns
15. Asian countries by clusterMost Indian food is in Clusters 1 and 3Sticky rice (Cluster 2) not common in India – shows preference for different type of riceMost countries are represented in every other cluster, indicating the great extent to which cultural cuisine is shared
16. Single vector decomposition of non-puddingsCharacteristics of 10 SVD Topics:Duck and seafood, beans, paellaPork and poultry; fermented foodsSweet and savory glutinous rice dishesSavory dishes with eggs, seafood and strong flavors of tomato, garlic, chiliRice noodles and rolls with meatPorridge and congee; simple staple foodsSweets with tropical flavorsStrong vegetal flavorsPungent vegetarian dishes, some dessertsRich foods: cake, deep-fried, cheese
17. Indian rice dishes(inc. puddings):Latent Class or SVD Clustering123TermScoreTerm ScoreTerm Scorebreakfast3.49seeds10.24sugar9.70vegetables2.97curry8.28milk8.75fish2.80leaves8.20slow-boiled6.58spice·2.17oil8.13sweet6.50seasoned2.10paste6.16dri·4.45boil·1.34garnish·6.16fruit·4.45fried1.32salt6.16nuts4.21lentil·0.81dal6.16jaggery4.21steam·0.39red6.16basmati2.12basmati0.21south4.15pongal1.73Simple foods with vegetables, and fish; especially breakfast foodsPungent curry and dalSweet dishes with milk, fruit, and/or nuts
18. JMP’s Text Explorer capable of telling the story of world cuisines in a quantitative wayPaints a picture of how people eatShows typical combinations and regional flavor preferencesMore diversity of rice dishes in a culture: they likely eat rice as a stapleThere are many ways to analyze rice dishes in this datasetWe just showed examples by type of food, by region, and within an individual countryCan be general or specific (as long as you have enough data)ConclusionsReferencesInspired by JMP blog story abour Chili Recipes by Scott Wise