Hacker News new | past | comments | ask | show | jobs | submit login
Mapping Neighborhood-Level Obesity in the United States (datainnovation.org)
68 points by sebg on Oct 9, 2015 | hide | past | favorite | 48 comments



Developer here, obesity is strongly correlated with race, ethnicity, educational attainment. So, yes, income and poverty play a key role, but are not the sole determinants of obesity. The obesity map doesn't look exactly like either a poverty map, or a race map, or an education map-- it's a synthesis of how those factors influence obesity. See the home page http://synthpopviewer.rti.org/obesity/ for additional background on how the data were created and be sure to look at layers depicting significant clustering of the results.


Thanks for that writeup link, that's really helpful in understanding what you've done. So everyone saying how the map correlates with various socio-economic factors is right (in various ways) because the obesity determination is entirely derived from socio-economic variables. That makes sense, but it wasn't clear until I read the writeup you posted that that's what we're looking at.


Here's a direct link to the tool instead of to an article about it: http://synthpopviewer.rti.org/obesity/viewer.html


I just ate a full meal from McDonalds this week for the first time in a while, and realized that I ate around 1000 calories, and was still a little hungry. This was really surprising, since 500 calories of healthy food can make me feel full.


Not that surprising tbh, Satiation isn't directly dependent on calories consumed but on various markers such as chewing, pressure on stomach wall and certain chemical responses.

Junkfood short circuits most of these, it's incredibly calorie dense and is easy to eat quickly since it requires little chewing, by the time your body signals satiation you can easily consume 1300 calories or more.

I've been on a calories restriced diet for the last 4 months (1500 or less a day) and I've eaten a few times at McDonalds, I just get a quarter pounder (higher meat to carb ratio, meat increases satiety more than carbs per gram) and a diet coke and then eat is in small bites and chew it thoroughly, by the time I'm done I feel the same satiation as I would have done bolting down a big mac and fries previously.

So far I'm down 45lbs with another 10 or so to go.


Assuming the full meal was a burger, there's basically no fiber. "Healthy food" tends to mean something including a bunch of vegetables, which are full of fiber


It's because of all the 'empty calories'.


Would be better to analyse other factors over the top of these. Income I guess is obvious, however transport probably is significant. Can you cycle/walk/get good public transport?

What about topography?


If you click through to the "SYNTHPOPVIEWER," you can see other factors on a similar map, but not overlaid. http://synthpopviewer.rti.org


My cursory glance seems to suggest higher obesity neighborhoods tend to be lower income ones. Though that observation is completely anecdotal based on my life experiences living in different parts of the country.


The USDA made a neighborhood-level map of "food deserts"-- communities where a significant portion of residents do not have easy access to large grocery stores and healthy food options. http://www.ers.usda.gov/data/fooddesert/ It would be interesting to see the two maps layered.


Yeah, that divide is extremely well delineated in DC, in an incredibly freaky way. Although I was quite interested to see the diversity of the pixelation in Great Falls which is one of the wealthiest areas of Fairfax County/ DC Suburbs. I suppose if you're really really rich you either take great pains to be in incredible shape or you just let yourself go completely.


I just noticed the same (and made a similar comment in response elsewhere).


I noticed the exact same thing!


It'd be easy plot income levels on the same map to see how the two correlate. It'd be interesting to plot other things like branches of chain restaurants, proximity to gyms, etc too.


I get the feeling they could simply interchange income maps for obesity maps and no one would know which is which.


So often these maps turn out to be population-density heatmaps. Do we know that is not the case here?

Ah, thanks nrjames, that page says its percent obesity within each grid cell. So not a population heatmap.


my anecdotal look at the bay area is inline with the grandparent comment, this seems to be basically a poverty map. I'm sure you could find places that has strange discrepancies where obesity and poverty don't directly correlate, but my guess is that in general they do pretty closely.


Great Falls, in Virginia (west of DC) is surprisingly obese, given it's an EXTREMELY wealthy area (avg home price is >$1 million) and also not racially diverse. Compare it with the nearby Herndon and Reston areas, both of which are very mixed, both economically and racially.

Interestingly, this pattern is the opposite of what you see when comparing NW DC with the rest of the city. NW DC is more white and wealthy, but slimmer. The rest of the city is relatively poorer and more diverse, and also more obese.


Thanks for the specific counterpoint! Definitely makes me want to dive into the data and identify other clusters that are atypical like that.


Poverty may correlate with historical cultural makeup. And culture historically correlates strongly with genes. And genes are largely responsible for obesity. Its complicated to be sure!


Genetics certainly contribute to obesity, but they aren't typically a primary factor, are they? Surely diet, exercise, and lifestyle are far more relevant. These are directly impacted by poverty.


In some population genetics is a strong driver of obesity. Pacific Islanders, Inuit, some Aboriginal America groups come to mind. Any isolated place experiencing periodic famine, supports genes to hoard calories.


Not exactly, Flushing in Queens is not obese, but isn't at all a rich neighborhood (assuming blue=good, red=bad, this thing doesn't have a key).

https://en.wikipedia.org/wiki/Flushing,_Queens


It's also full of east asian immigrants, which are less likely to be fat (look up the obesity rate in Korea, for instance - it's almost 0)


Click the "Info" tab near the top for a key.


I wonder if parental income maps would be an even better match. I've heard that when one is raised as an obese child, the chance of reaching a healthy weight later in life is very slim.


In rural areas, the neighborhood squares cover individual houses. I'm seeing some where my daughter lives that cover individual people. They seem accurate, the blues and reds live next door to each other.


In Los Angeles, it maps out to ethnicity pretty nicely. I don't know why, I wouldn't think it would, but it clearly does.


http://www.cdc.gov/obesity/data/adult.html

"Non-Hispanic blacks have the highest age-adjusted rates of obesity (47.8%) followed by Hispanics (42.5%), non-Hispanic whites (32.6%), and non-Hispanic Asians (10.8%)"


It maps out ethnicity in nearly every metro and even suburban area I can see.


Yeah, even more so than income. It is kinda weird that way.


The only thing I can think of is vestigial cultural rules being a determining factor; what you eat, when, how much - what your hobbies tend to be, how much you move in life.

Every individual of course sets his or her own path, but populations taken as a whole seem to exhibit dominant patterns.


So the only thing you can think of to explain this is cultural factor?

You don't think genetics could be a factor here at all?


Sorry. I claim zero knowledge of epidemiological obesity studies or theories. In retrospect, it was not my place to conjecture at all --- I really don't know anything about this.

Although for a thoughtful and thorough analysis:

An introductory wikipedia article: https://en.wikipedia.org/wiki/Epidemiology_of_obesity

Google scholar search: https://scholar.google.com/scholar?q=epidemiological+obesity


Interesting. My read on this map is that it also tracks the urban/non-urban breakdown very neatly: as you get out of the urban code, obesity %s increase dramatically.


I noticed something similar as well, especially in Houston, where the eastern side is much more urban, and the western side is where the expansion has been and is less urban and contains more neighborhoods, smaller grocery stores, etc.


Interesting - typically I would associate "urban" with smaller grocery stores.

i.e. there are 8 grocery stores within a few minutes walk, but they are all smaller - some just sell produce, others are smaller versions of suburban grocery stores.


There seems to be some kind of 'periphery' effect too, in places: zoom in on Portland, Oregon, and it sort of looks like there's a red ring around the town.

I could see it being a 'rural' thing, but the ring is weird.


That could just be from gentrification pushing lower income people out of the city center. Similar things have happened in DC, Chicago, etc. This Wired article shows some of that effect. http://www.wired.com/2013/08/how-segregated-is-your-city-thi...


I don't think so, at least in Seattle. Looking at the city breakdown, it corresponds to transit coverage and (my perception) transit friendliness. It actually seems to correspond inversely against income, since the transit coverage is vaguely inverse with Seattle income.

This is, by the way, a strong conflict with East Coast cities in the US, where there's a decided "obese area" visible. I don't have insight into those from the ground level.


It might be effect of people who have to drive for most trips, versus those who can walk or bike (alone or in combination with transit) to at least a few destinations.

That sort of access also makes those places more desirable, and thus more expensive, and thus more likely to be populated by rich people.


You can see it all over the Western us. Super fascinating.


Walking turns out to be good for you.


This map follows the income distribution in my area very closely. Neat, but sad.


How do we fix food deserts? I have no idea but I wish we could solve it.

If I was a billionaire I would subsidize running Whole Foods in impoverished neighborhoods.


A big misconception about fixing food deserts is that if we could "just add healthy options", the problem would be solved. But it's not that simple ~ it's an education and cultural issue. Where I live has a lot of red dots on this map and is not considered a food desert, I've got several grocery options. The issue is folks here are deeply steeped in modern southern food culture (fried on top of fried everything. all white bread and meat. very few vegetables, if any) combined with low income and education. They simply don't want to eat "rabbit food" (I'm quoting from actual conversations). Suggest anything different to them and they are offended. My (obese with numerous health issues) neighbors think my wife and I are "libruls" because we eat normally. It's a whole different mindset that you must empathize with and understand before talking about how to fix it.


Where does the data come from? It doesn't say.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

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

Search: