diff --git a/writing/report/project_report.tex b/writing/report/project_report.tex
index d158a53..d8cbeb2 100644
--- a/writing/report/project_report.tex
+++ b/writing/report/project_report.tex
@@ -118,8 +118,14 @@ sorting=nyt
     \section{Discussion}
     \indent
 
-    TODO
+    To try to answer our research questions, we made 3 user samples: \texttt{500-pop}, composed of the 500 most popular users on Twitter, \texttt{500-rand}, 500 random users on Twitter, and \texttt{eng-news}, the top 100 English news channels on Twitter (3).
+
+    Our first research question asks: \textbf{how frequently do people post about COVID-related issues, and how interested are people to see COVID-related posts?}
+    After making graphs showing the frequency of COVID-related posts, the histograms showing frequency all tend to be skewed right, or decreasing. In all 3 samples of \texttt{500-pop}, \texttt{500-rand}, and \texttt{eng-news}, for every user, the majority of their posts were not COVID-related. However, news channels tended to post more about COVID than \texttt{500-pop}, who tended to post more about COVID than \texttt{500-rand}. This suggests that the \texttt{500-pop} users tend to post more than the average person about COVID, possibly because they have a large influence and want to share their opinion, but they post less about COVID than news channels, who have to report the latest news about the pandemic, regulations, and slowing the spread.
+    The histograms showing the popularity ratios are also skewed right, except the news, which is almost centered. This shows that for COVID-posts outside of news channels, they usually receive poor engagement. However, for news channels, they receive almost as much engagement as their other posts, suggesting that users are indifferent to COVID-news as regular news.
     
+    Our second research question tackled the same thing: but instead, how it changed over time. From the line graphs showing COVID-posting frequency over time, it tends to decrease for the samples of \texttt{500-rand} and \texttt{eng-news}, although \texttt{500-pop} and \texttt{eng-news} shared a peak around December 2020 and trough around June 2021, which seems to be related to the rise and fall of COVID cases in the US. A bit surprisingly, the news has stopped posting as many COVID-related posts, maybe because they are starting to incorporate other keywords such as "Delta variant". In addition, there is a large spike in May 2021 for the sample \texttt{500-pop}, which is also around when the Delta variant started appearing. This could mean that the news started talking about that more, while the popular twitter users talked about its consequences. However, the \texttt{500-rand} sample has seen a decrease in frequency, suggesting that they post less about it, possibly because their COVID posts get less engagement, or just a general decrease in the interest in COVID as a topic. We speculate it can possibly be people getting used to seeing COVID-related news all the time, and just like background noise, are starting to ignore it (a process called habituation).
+    The line graphs showing popularity ratio over time, however, were a lot messier. Despite attempts to filter out the noise, there were constants spikes in the graph throughout the entire graph, making it difficult to draw any conclusions. This could possibly be because the sample size is too small, or the method of determining popularity isn't sophisticated enough and prone to random noise. Unfortunately, this means that further research is required to answer our question.    
     
 \nocite{*}
 \printbibliography