For this week’s blog post, I decided to create a data visualization for body fat. The data contains several statistics related to the measurements from 252 men including their total body fat, age, weight, height, neck size, wrist size, thigh width, and other statistics. I was particularly interested in seeing whether or not certain factors contributed or were an indication of body fat. I utilized a scatter plot in order to visualize the most common trends and determine whether or not there would be a correlation between values on the x and y axes. The first visualization I created compared age and body fat. I believed that as age increased, so would body fat. To my surprise, there was no direct correlation between body fat and age, as evidenced by the visualization.
Next, I tested thigh width, which appeared to be positively correlated with increasing body fat. Through this test I became curious as to whether or not other areas of the body gave such a strong indication of body fat. Interestingly, when comparing body fat and wrist size, the scatter plot demonstrated that generally, as wrist size grew, so did body fat. Overall, Google Fusion Tables was a very powerful tool when studying general trends that utilized quantitative measurements. It was really easy to switch between different groups of data by simply clicking on different columns in my data sheet. Unfortunately, while the scatter plots are great at demonstrating group trends, individual outliers are not effectively represented and therefore, when addressing causality, one cannot say that any of these factors is a true indication of body fat (or any other correlated data set).
Very nice job, Ethan, both on the scatterplots and the analysis.
Your data visualizations are very well done and provide a good representation of the data provided by the dataset. I do wish that you included labels on the x- and y-axes of the first scatterplot, but it is somewhat of a moot point because the chart is very intuitive in itself. You did a great job in elaborating on them and providing background information for the dataset in general. Your critique of Google fusion tables at the end is also a good addition.
I really liked how your post separated scatterplots by different regions of the body as I thought this was very informative. I also enjoyed your commentary on the accessibility and user capability of the google fusion tables.
I really like how you have organized your blog post, especially with two photos that reveals new meaning to the table. I like how you wrote the reasons why you were interested in this topic!
Some thing I would have liked to know from the project is what exactly was wrong with the original data in table form that didn’t allow us to understand the hidden meaning from the data.
Your scatterplots were really easy to interpret while also being really indicative of your commentary on the data. The organization of your analysis was straightforward and easy to understand, and I like how you showed us the ways in which you went about making sense of the data in the first place. I would have never automatically thought that wrist size would be so suggestive to body fat!
Your representation of the data through individual scatterplots comparing body fat with thigh width and wrist size is very helpful in determining the correlation between specific body parts. Often times, a dataset presenting several statistics can be overwhelming because there are so many factors that arise among relationships of different values, but your choice of graph and analysis did a great job honing in on the specifics.
The representations of the data that showed the correlation between thigh size/wrist size and body fat are interesting, and very easy to read! It would be interesting to see if you could put a trend line in there to see how strong the association is, because that would definitely add another dimension to this data. Your explanation for it was very clear and concise, especially because you honed in on the correlations of thigh and wrist size to body fat. (Also, fun fact: growth hormone, testosterone, and estrogen levels are determining factors for fat growth on different parts of the body, as well as how fat develops overall)
Great job on the visualization! It’s very clear and easy to read. I’m kind of curious what the margin for error might be on these, and I’m glad that you addressed the issues of correlation and causation in your dataset
Great job with using visualizations to answer some of those hypotheses that you had. I really liked how you would present your initial hypothesis and then showcase the visualization and your analysis right afterwords.
A minor critique that I would include would be to include axis labels as well as possibly a trend line. Including the labels helps your chart stand out more on it’s own while the trendline helps a viewer quickly determine at a glance if there is a strong correlation or not.
By using the scatter plot you truly demonstrated the correlation between body fat and size of certain parts of the body. It is a wise selection of the visualization method, combined with creative hypothesis and clear analysis. I totally agree with your interpretation on the undetermined causality. I believe that if the radius of the dots can be smaller, the visualization will appear even more attractive.
Nice data visualization, they are easily readable and represent clearly the data you used. In terms of the relationship between wrist width to body fat, I somewhat struggled trying to figure out why this would be useful information, and how it may be beneficial. Not that I think it’s useless information but perhaps the data visualization could convey a more causal relationship