Russian Ukraine war on tweeter

Inspiration

Since Russian-Ukrainian war is one of the most controversial issue on the globe, we wanted to analyze the public opinion regarding such issue through social medias with high usage. Although it is is war between two countries, we wanted to conduct a global public poll since it has a high influence globally.

What it does

It retrieves Twitter posts using twitter API and processes the data by conducting several filtering labels. Then it filters tweets that are only related to Russian-Ukraine war, and conducts a sentiment analysis on the given data.

How we built it

Collected twitter data via tweepy library, then preprocessed the data using various filtering methods. After that we conducted a sentiment analysis and trained the model after vectorizing. Then we developed a website and visualized meaningful data.

Challenges we ran into

There were cases where negative statements were classified as positive statements due to wordings of the posts. The sentiment analysis algorithm we used classified statements solely based on unigram model. When using unigram model, the algorithm wasn't able to correctly classify several tweets. For example, tweets including word "support" were identified as positive, since the word itself implies positiveness while some of them should be classified as negative response because "support Ukraine" is not a positive response towards the war. To overcome the issue, we created custom labels that would correctly classify the tweets. We trained the model with custom labels to correctly classify tweets.

Accomplishments that we're proud of

We are proud of our web development, because none of our teammates had experience with web programming before. Also, we are proud of the fact that we trained our own model to remove misclassified tweets.

What we learned

We learned how to process data effectively, how to use tweepy, how to sentiment analysis, how to analyze&visualize data, and mostly how to cooperate with teammates. In terms of the topic, we were able to feel that the war hurts people.

What's next for Russian Ukraine war on tweeter

Receive a geographical information on tweets to analyze the data by different countries.
Website Live update
Using our sentiment analysis on different topics