The first plot I created was a parallel categories plot showing the relationship between voting frequency,
income level, and education level. This plot shows the increase of voting - both sporadic and always - as income and education levels
increase. The percentage of high school graduates or less who are sporadic or always voters is 64.8%. This number increases to 76.8% after
some college is completed, and 81.9% after college graduation. Based on this plot, it looks like increased education levels lead to
increased voting.
This plot also shows income level of these voters. Most college graduates earn higher salaries than those who
did not graduate college or continue with their education after high school. People who have a high school degree or less tend to earn
less than $40k. And the people who did not finish their college degree are sort of equally split between the four income brackets - 26.3% earn
less than $40k, 26.8% earn between $40-75k, 29.8% earn between $75-125k, and 16.9% earn more than $125k. This plot shows how an increased level
of education can lead to an increased income, and as a result, those people with a higher education and higher income are more likely to vote
either sporadically or all of the time.
The second plot I created was a bubble/scatter plot. This plot has, on the x-axis, the percent of (in person) voter turnout in
elections from 2014-2022, and the y-axis shows the percent of absentee votes in those elections as well. Each point represents a county
in Virginia, and the size of the bubble represents the total number of registered voters in that county. This plot is interactive as well,
and viewers can scroll back and forth between the years to see the differences between them.
Something to note about this specific data is that there were some percentages that were calculated that were reaching infinity due
to lack of data, and others that were greater than 100%. The rows containing the infinite datapoints were removed from this model.
The values that were greater than 100% were not shown in the graph, but they are real data that were collected, so I did not remove them
from the dataframe.
With that said, this plot is very interesting. Most of the data from 2014-2018 remained at the same level, but in 2020 there was a huge jump in
absentee voting. There was also an increase in in person voter turnout, but this could have been due to the fact that it was a presidential
election year (2016 also had an increase in person voter turnout). The rise in absentee voting in 2020 could have been because of the COVID-19 virus,
which forced many people to stay home from the polls and vote through absentee ballot. Additionally, it was an election of historic proportions
between Donald Trump and Joe Biden, with increased movements to try and get people to vote, either through in person or absentee voting.
The third plot I created was a plot showing the percent of voter turnout in different counties and cities in Virginia in the 2022 elections. This plot has the benefit of seeing the trends through color upon first glance. As shown in this plot, counties in the southwest of Virginia had lower in person turnout rates than counties in the northeast part of Virginia. This is supported by the popup I added, which shows the exact percent turnout for that area when you click on each county/city. Based on my plot, it looks like geographical location is a factor in in-person voter turnout.
Besides adding the popup in the third plot, I added color and changed the font style of the font on this website. I also padded the text so that it did not appear to run off the edge of the page. The additional color and different font makes my page more fun to read, and less like a scientific paper! I also increased the bin size of the colorscale in the choropleth map to make the distinctions between levels easier to see. There were some troubles that I ran into - particularly with the popup. I had trouble figuring out how to modify the json file, but once I was able to get that working, it definitely added to the effectiveness of my plot. With the popup, viewers could see the exact percentages instead of an approximation based on the colorscale.