Hacker News new | past | comments | ask | show | jobs | submit login
Show HN: Intuitive nutrition information (spe.lt)
39 points by gusgordon 9 months ago | hide | past | favorite | 40 comments
Hey everyone, I've been building this nutrition tracker and calorie counter recently, after being frustrated by existing products for ages. I built a similar app 8 years ago [1], but came back to this problem again since there are still no good solutions here. Lmk your thoughts and improvement ideas :)

[1] https://news.ycombinator.com/item?id=10077618




Have to comment since I've been listening to Dr. Lustig lately (https://youtu.be/dBnniua6-oM?si=6mS-3SKI7R4zOdYT).

The site shows that 1 can of Coca-Cola has fewer calories and less carbs than a medium apple: 90 vs 122 calories, 25 vs 30 gram of carbs. There is no other information, e.g no fiber content or added sugars section. Would this lead someone to believe, against common sense, that Coke is equivalent or even a healthier choice?

Edit: For those interested, here is a link to a similar system that contains more information (albeit not as polished): https://perfact.co:8443/Nova/


Adding "fiber" in the results (and changing "carbs" to "net carbs") would help.

I never understood the motive of highlighting "added sugar" though. Who cares if it's considered "added" or not? I recently was browsing the cranberry juice section at a grocery store, and without going into detail, it was obvious that some brands are gaming the definition of "added". Also, many natural sugar sources have unhealthy levels of sugar. Perhaps a metric like sugar density or caloric density would be more useful.


It’s a successful propaganda campaign. There are many people that wrongly believe that “natural sugars” are somehow better/more nutritious. (My mother is one, and eats way too much fruit and fruit juice because “it’s natural!”)


natural sugars come with free vitamins and minerals.

but sure, you can have too much natural sugars.


Yep. Any site do this instantly lost credibility to me. Total carbs is total carbs. If they can't do basic cico it's misleading


There’s a lot of nutrition label gaming going on these days. It’s my understanding that the nutrition label is mostly enforced on the honor system so, probably easy to game.


Added sugar doesn't have as many nutrients, and is absorbed by the body much more quickly.


this is not true, sugar is sugar* -- what matters is how it is consumed, not how it is added. if sugar is consumed with a lot of fibers then it will be absorbed slower by your intenstines. the same fruit consumed as juice or whole fruit impacts how the sugar is absorbed.

[1] sugar is actually multiple sugars, and different sugars (glucose vs fructose vs sucrose) - https://www.healthline.com/nutrition/sucrose-glucose-fructos...


But, in general, processed sugar has the effects I mentioned, because it is added.


> Added sugar doesn't have as many nutrients

Most overweight people (the audience of a calorie counting app) don't need more nutrients, they need less calories. Besides, fruit is a really poor source of vitamins/minerals in general, so this is a weak ground to stand on.

> [added sugar is] absorbed by the body much more quickly.

You're using circular logic.


While you certainly succeeded in simplifying the UX, I'm not for simplifying the data on food. Is there any database of "concentrations of vitamins, minerals, and micronutrients" you could integrate?

I know this would defeat the purpose of a "simple" tracker, but I believe that we need to support the message that food is "complex" [1].

[1] https://www.bionutrient.org/bionutrientmeter


Totally - right now it's using the USDA food nutrient database [1] which has all nutrients, so all the nutrients on the back-end are there. Calories is the most important thing to the majority of people (in my experience), so there's a balance between showing too much information vs. keeping things simple. Probably should just add a setting to show everything, though.

[1] https://fdc.nal.usda.gov/download-datasets.html


I entered "Cheerios" and was told one box of Cheerios cereal is 61 calories; and that one box of blueberry Cheerios is 611 calories. So I think that we need better quantification of units than just "box," since presumably these measurements are dealing with boxes of different sizes.


Man, those GMO blueberries are really getting out of hand, aren't they?


It's 2023 and we don't have GMO blueberries the size of a basketball yet? Lazy ass scientists need to get back to work.


Similarly, 1 cup unsalted peanuts is listed as having 10x the calories of 1 cup peanuts. The latter probably refers to peanuts in their skins while the former refers only to the nuts which would be packed much denser. A photo of the food could help here.


Seems cool! Congrats!

I agree that most products - I stumbled upon and tested - in this space are frustrating.

Would be nice to have a good UX to split food by meals (and have subtotals). But maybe that is too complex and you were not aiming for that!

Additional feedback: I think knowing if the food is weighted raw or cooked is pretty important. Sorry if I missed something but I couldn't tell if the "whole wheat pasta" were cooked or not. Uncooked was available for rice, though.


That's cool, but it seems to be limited to macros and working under the assumption that a calorie is a calorie.

It would be much more interesting to me to have an easy way to access the information about "how processed" a food is, for example using the NOVA classification[1,2,3] system or something like that. I'm not sure if you can access that as a DB download or an API of some sort, but worth looking into.

If you can't get data on this, then I'm pretty sure you can approximate the nova score by based on the "added sugars" and fibre contents... although it would be difficult to tell apart "undisturbed fibre" vs. fibre added later on as additive.

[1] https://en.wikipedia.org/wiki/Nova_classification [2] https://www.fao.org/3/ca5644en/ca5644en.pdf [3] https://world.openfoodfacts.org/nova


Why the hell would a person need an app to tell "how processed" a food is? This is the one bit of information that is contextually obvious.


It's just a nuance to the rule of thumb, that doesn't warrant such a strong reaction?

If a 1st order rule of thumb is keeping caloric contents in check, a 2nd order rule doing that for nutrients individually, I wonder if someone knows a good but easy third order rule of thumb that takes into account processing.

Something like "when two foods have roughly identical nutrients, prefer the one that has ingredients you could buy or make at home" might either not help cause there are no two such options to choose from, or it's too obvious (fresh pizza at the supermarket is worse than a handmade pizza)

Is there a good rule of thumb that helps you pick relatively healthy foods in a world where many foods are ultra-processed already, and can't really be avoided?


My 1st rule of thumb: Avoid processed foods. This way I don't have to really think much more about what to eat.

And yes, ultra-processed foods are easy to avoid if you go the right suppliers, such as grocers. Supermarkeups are definitely difficult places to eat well.


Consider it for how processed a specific brand of food is, rather than a general food. Right now most specifications don’t get more specific than organic, free range, grass fed, vegan, all natural, no preservatives, etc. I suspect the problem would lie in data quality more than deriving insights.

For example: a sliced apple is technically processed food, but what is it sprayed with?


Usually hydrogen peroxide[0] (when sliced) but if you are referring to what chemicals a farm uses, that would be really, really hard to provide. I believe Chipotle tried to do a farm to store supply chain doc and gave up and went with a distributor. However, they still are trying w/ robotics and other initiatives.[1]

At the end of the day, how do you really know what chemicals are used on the farm? The cost of tracking anything is astronomical. I.E. cost of a bolt in airline industry.

Lastly, at least for food: producers, distributors, end-users (stores), and everyone else uses different software with proprietary protocols and the margins don't really incentivize anyone from switching.

[0]: https://www.sciencedirect.com/science/article/pii/S0362028X2... [1]: https://www.supplychainbrain.com/articles/35217-chipotle-zer...


Exactly my point, thanks for elaborating the example!


Different organic certifications allow different inputs, look up e.g. NASAA https://www.agriculture.gov.au/sites/default/files/documents... but wow is that a difficult document to apply when shopping!


I don't know, it's not really easy for me (as an ignorant American who grew up on junk food) to guess how much processing is in food. I dunno if there's much difference between cow milk, soy milk, oat milk, etc. How is tofu or blueberry yogurt made? What's in a slice of sandwich cheese? Are these fancy organic crackers any different from Ritz? What's in olive oil vs canola oil? Vegan butter vs margarine?

Most of the grocery store is made up of processed flavored glops of one sort or another. The Nova classification is new to me, but would certainly be interesting information (though possibly also fearmongering).


> I dunno if there's much difference between cow milk, soy milk, oat milk

Well given the fact that cow milk exists in nature, has been drank by humans for thousands of years, and the other two are trying to be like cow milk, right off the bat, I think a good working hypothesis is that cow milk is less processed than soy and oat milk. They may pasteurize and homogenize it and add some vitamins, but it is fundamentally pretty close to what came out of the cow. Soy mil and oat milk, need a lot more processing to turn soy and oats into something at resembles milk.


Isn't oat milk just putting oats and water in a blender basically?


That's the exact question I had too. I don't know they just add water and soak, or if they use some sort of chemical process to dissolve the oats and isolate some powder that they reconstitute later... etc.

I feel like this should be information available on all labeling. I have no idea what "guar gum" is, and whether that's more processed than "enriched wheat flour", and whatever the hell "natural flavors" are.


Chris van Tulleken has an excellent book (and talk [1]) on the subject

[1] https://youtu.be/5QOTBreQaIk


This is a really phenomenal talk (and he's a great speaker)! Thank you so much for sharing this.

I haven't studied this topic much since the days of Michael Pollan's In Defense of Food, and I really appreciated Chris's scientifically grounded overview. I think I'm going to get the book and seriously reexamine my eating habits in this new year.

Thank you again for linking this!


You have just asked all the right questions. Now comes the fun part of looking up all the answers :)

Spoiler: Keep it simple and learn to cook.


I’m a big fan of MacroFactor. It’s one of my few subscriptions. It’s the only tracker that works for me. The developers make understanding nutrition as simple as possible without pretending it can be oversimplified.


How about adding our own data ?


Unrelated to the actual product, the kerning in the wordmark needs work.


Cronometer exists already and does this a million times better


Didn't know Chronometer so I checked: you have to sign up.

This already makes OP project better for me (subjective of course)


Great interface, What source(s) are you getting the data from?


Answer available in a comment from OP: https://news.ycombinator.com/item?id=38810773


Nice! Howabout adding a way to link to a meal?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: