# Data Visualization and Analysis, Part 2/3 – United States Department of Agriculture

By | February 24, 2016

Author: Qi Chen

# 1. Introduction

In the previous part, we introduced some data visualization techniques and showed the relationship of GDP, healthcare investment and life expectancy grouped by countries. In this article, we will focus on poverty and obesity, and analyze the dataset from United States Department of Agriculture. Specifically, by analyzing the dataset using data visualization techniques, we want to explore the following questions:

1. How does poverty rate vary within racial and ethnic groups?
2. Does the milk/soda price ratio influence obesity rate?
3. Are the obesity and diabetes rates higher in fast paced states like New York States than in agricultural states like Wisconsin?

# 2. Analysis

## 2.1 Software used

In this article, we will demonstrate the use of Tableau for data visualization.

## 2.2 Approaches

In order to explore the relation between poverty and ethnic groups, we rank the poverty rate from highest to lowest for different counties. In Graph 1, the x-axis represents the FIPS (Federal Information Processing Standard) of each county, which is a unique numeric code, and for the y-axis, we display the percentage of different population groups for the county. The y-axis variables include:

POVRATE10:
Poverty rate of counties in 2010

PCT_NHWHITE10, PCT_NHBLACK10, PCT_NHASIAN10, PCT_NHPI10, PCT_NHNA10:
Percent of different racial and ethnic groups in 2010

From Graph 1, we find that poverty rate is relatively higher in black and native American people.

Graph 1: Ethnic groups vs. poverty rate

Then we try to discover the relation between milk/soda price and obesity rate. In Graph 2, the x-axis represents the FIPS of each county and y-axis represents the milk/soda price ratio. The obesity rate for each county is represented by different colors. With the color closer to red, the county has a higher rate of obesity. Graph 2 clearly shows that counties on the left are more likely to be red, which means higher milk/soda price ratio is associated with higher obesity rate.

Graph 2: Milk/soda price ratio vs. obesity rate

In Graph 3 and Graph 4, we show the obesity rate and diabetes rate for each state in the map respectively. Red means a higher rate of obesity or diabetes, and green means a lower rate. From the two graphs, we can see that generally, the two rates are positively correlated. States on the east are more likely to have higher obesity rates and diabetes rates.

Graph 3: Obesity Rate in U.S

Graph 4: Diabetes Rate in U.S.

## 2.3 Results and Summary

In this demonstration article, we explore the poverty rate with respect to different population groups, as well as the obesity and diabetes rates for different states.

This article is based on a course project in Industrial Data Analytics offered by Prof. Kaibo Liu in University of Wisconsin-Madison in Spring 2015. I would like to thank Prof. Liu for his instruction and also thank my teammate Li Xie and Sowmya Shankar for their initial work.

# 3. Source

http://www.ars.usda.gov/Services/docs.htm?docid=8964

•
•
•
•

## 9 thoughts on “Data Visualization and Analysis, Part 2/3 – United States Department of Agriculture”

1. Annie

it would be very interesting to additionally look at ethnicity vs obesity-this data has brought a lot of questions to mind! will there be a follow up piece comparing race and health metrics?

2. Michael Baer

I agree with Annie that it would be interesting to look at comparisons of ethnicity, obesity, and geographic distribution to help link all the data together to create a better analysis. From simply comparing the multiple graphs, it is clear there is an association between obesity and diabetes rates however the soda/milk ratio and ethnicity data seems unrelated to the rest of the set.

3. Erik Pechnick

Recently I know there has been a nationwide soda tax to try and help with the obesity issue. It would be interesting to see a before/after with this tax to see if it has helped at all. Also, It would be interesting to link obesity with population density to see a correlation between obesity and big cities

4. Eric Fleming

I am curious what factors impact the milk/soda ratio, besides perhaps proximity to agricultural land. I’m also very curious to hear about how this metric was chosen to be studied, I’ve never heard it discussed before, but it does seem to have some correlation! However, I will add that the graph relating to milk/soda price was difficult to read and interpret for me.

5. Dylan Weber

I am confused as to why you used milk/soda price. In my opinion, these are two completely different metrics and would show different results. Wondering what your thoughts are on this and the reason behind you doing so?

6. Lauren Chiang

Thank you for sharing your analysis. I found it interesting that there was a correlation between the milk/soda price and the obesity rate. However, upon further thought, it could be attributed to the fact that with a higher price of soda, other foods may also be expensive and therefore healthy choices may be pricey. It was difficult to interpret the graph with the ethnicity and poverty rates. I did like the graph comparing diabetes and obesity rates as it was easy to see the correlation.

7. David Wilkins

One comment I do have is that the first figure is difficult to read, which makes the point you’re trying to get across hard to understand. I assume your goal of using milk/soda price was to measure the general ratio of healthy food offerings to unhealthy food prices; however I’m curious to see if there’s a more standardized ratio across a greater variety of food offerings. The conclusion from the graph does make me think about another analysis on this website where there seems to be a correlation between accessibility of healthy offerings, and more healthy people. This means that people are happy to eat healthy foods when they’re economically viable.

8. Pooja Shivale Patil

The correlation between soda price and obesity is a bit surprising as we have a strong opinion on those being negatively correlated. Also, this one supports analysis which was done recently where obesity rate on income levels was assessed.

This site uses Akismet to reduce spam. Learn how your comment data is processed.