database design - food information graph modeling -

i've following information mongodb db want export neo4j graph:

"_id" : 1366, "fields" : { "title" : "beef steak", "keywords" : [     "beef",     "steak" ], "categories" : {     "category" : "meat",     "subcategory" : null,     "verified" : false,     "predicted" : true  },  "nutrients" :     {         "tag_name" : "fasat",         "value" : 1.352,         "unit" : "g",         "title" : "saturated fatty acids"     },     ... 

i've been reading neo4j modelling examples i'm not sure model.

this i'm thinking:

nodes:

• food
• keyword
• category
• subcategory
• nutrient

relationships:

1. food - rel_name -> keyword
2. food - rel_name -> category
3. food - rel_name -> subcategory
4. category - rel_name -> subcategory
5. food - has_nutrients -> nutrient

the "has_nutrients" relationship name contain "value" property indicate how of nutrient food has.

i thought 2,3 , 4 relationships have possibility perform different queries: - subcategories of category 'meat', - category , subcategory of food, -all foods 'pasta' category. maybe there's more simple way.

if can't think of better names relations,

food - [has_keyword] -> keyword food - [has_subcategory] -> subcategory subcategory - [belongs_to] -> category food - [has_nutrient] -> nutrient 

for subcategory, typically direction child parent, that's more preference.

will need differentiate between category , subcategory? if category meat , subcategory pork, ever need link meat? have connection pork , imply meat cut down on relationships.

({name='pork', type='subcategory'}) -[belongs_to]-> ({name='meat', type='category'}) ({name='pork chops', type='food'}) -[has_subcategory]-> ({name='pork', type='subcategory'}) 

this cut down on relationships, you'll able imply category subcategory.

aside above recommendations, rest looks graph oriented , looks solid design.