1. Objective

After the release of a new game, it is important to know whether it has been accepted by the community and to know if there are some issues that should be adressed.

A popular MMORPG, Guild Wars 2, has recently released its third expansion, "End Of Dragons".

Using the Twitter API to collect data (using the "#GW2EOD" hashtag), and BERT to estimate sentiment, we can get an idea of how has the game being received by the community.

2. Results

From the conducted analysis, we can that overall, the game seems a success and the community seem to appreciate and enjoy the new release by ArenaNet.

• The vast majority of tweets seem to be very positive and some of the common words to describe the game include: "fun", "beautiful" and "amazing".
• Nonetheless, among the negative tweets, players seemed to complain about a certain meta event (important for the players), that contained bugs which made it more difficult.
• This has been adressed by the game developer and it is possible that the number of negative tweets could decrease on a future analysis.

3. Overview

# Preprocessing

• Collected data using Tweepy and the Twitter API
• nstantiated a BERT model, "finetuned to sentiment analysis on product reviews that predicts the sentiment of the review as a number of stars (between 1 and 5)" (more details)
• Cleaned the dataset using regex and the Re library

# Modeling

• Applied the model to the clean dataset to obtain the sentiment of each tweet
• Plotted cloud of words from negative tweets (sentiment = 1) & from positive tweets
  (sentiment = 5)
• Using TextBlob, got the subjectivity and the polarity of each tweet

Some images from the project