Introduction

One of the major aspects of data analysis is visual interpretation. In this project, we utilized the FBI data for arrests in the United States in 2012 by sex (Table 42). Using this data, we did our best to create four different graphs: bar charts, pie charts, box and whiskers, and line graphs. Through this process, we both analyzed the best graphs for interpreting the arrest rate data and explored the flaws with the other graphs. 

To do so, we analyzed the arrest data based on gender. For each type of graph, we used Tableau to create a visual for both sexes. Then, we examined the effectiveness of each graph solely based on our abilities to analyze the data for these visuals. 

Bar Chart

Pie Chart

Box and Whiskers Plot

The range of arrests of women per category in America is around 480 thousand which is much less than the men data coming in with a range of around 950 thousand. This shows how out of all arrests in 2012 males had a much higher ceiling of number of arrests in one single category.

We found that there are very significant outliers in that data. In the women's arrests there are two outliers — by eye — sitting around 440 thousand and 480 thousand arrests. In the men’s arrests there is only one outlier at 950 thousand. This is important as it throws off our analysis which can already be seen in our spread greatly inflating our maximums.

The median of women arrests is around 20 thousand while the median of men arrests in a single category is around 70 thousand. This gives us a decent idea of where the average data falls as medians are more resistant to the outliers covered previously. With this information we can speculate that per category men have more arrests than women.

Both data sets — men and women — are heavily skewed right which shows how the number of arrests in many categories are more concentrated at lower numbers while only a few categories have a relatively large amount of arrests, this is true for both men and women. This is important as we can speculate that a majority of categories have a small number of arrests while only a few have a very large amount of arrests.

While the box and whiskers plot is possible, due to the large range for the data, there are definitely more viable options.

Line Graph

While it is possible to create a line graph using time vs. the number of crimes committed, we were unable to find data with multiple dates over time in 2012. As a result, we are unable to create a line graph for the trends overtime regarding crime rates in the United States during 2012. However, if we are to graph data over time, we would create two graphs based on sex and create individual lines based on offense charged.

Conclusion

Of the graphs we created, the most effective graph was the bar chart. While the pie chart and box and whiskers plot were both possible, they were not best suited for the given data. In addition, the line graph was not possible using the given data.

Through our usage of these graphs, we were able to determine that, while it is possible to create and analyze multiple types of graphs, not all of the graphs suit the given data. As such, one should not decide what type of graph they will use prior to obtaining the data. In order to create an effective visual interpretation of the data, one must determine the most well-suited type of graph based on the data provided. 

Works Cited:

“Table 42.” FBI, ucr.fbi.gov/crime-in-the-u.s/2012/crime-in-the-u.s.-2012/tables/42tabledatadecoverviewpdf/table_42_arrests_by_sex_2012.xls.