This summer has allowed me to use my current skill set of basic text analysis while also improving on my flexibility and open-mindedness in research. While I have made progress on my research project, and accomplished many of the goals that I set out with at the beginning of the summer, I understand that there is still more work to be done.

Over the next few weeks, I will be working on displaying and showing my results. I have found the major themes of the misinformative tweets and found certain themes that are not as prominent. Certain themes, such as “anti-vaccine” and “government oppression” were prominent, while other themes such as “climate change isn’t real” were less prominent. Overall, I have found that there is a definitive relationship between time and prominence of other themes. However, there are a few themes that are also constant and consistent over time. I am excited to present these findings.

I have begun to use Tableau to generate the visualizations and will use that for the timeline as well. Because of my prior experience with Tableau, I only required a refresher course. Hopefully this data will be even more accessible to information seekers with the visualizations that I have created. I hope to place them on my personal website.

I have also begun to work on the paper component for this research. Discussing my findings and placing them within the context of current misinformation research will show the importance of viral misinformation. Furthermore, as we seek to identify how misinformation spreads, misinformative handles such as the one I used will prove invaluable. I am still unsure which platform to place the finished paper on (journal, open access archive, etc.) However, I would like to share the findings.

One major struggle was the medium I chose to use. Over the last year, Twitter as a platform has become harder to access without a verified account. Obtaining a verified account is also substantially more expensive than it was a year ago. As a result, although I was able to avoid a worst-case scenario of being unable to use any of the tweets or images, identifying and revisiting tweets online proved to be difficult. I chose to avoid certain tweets that contained only images with no context simply because it was so challenging to revisit the tweets (even with the Wayback machine.) However, this only affected a relatively small number of tweets, and I am still able to present text-only tweets as well as tweets that contained an image with a description.