The Electronic Entertainment Expo 2018 (E3 2018) was the 24th of its kind, during which hardware manufacturers, software developers, and publishers from the video game industry presented new and upcoming products to attendees who were primarily retailers and members of the video game press. The event, organized by the Entertainment Software Association (ESA), took place from June 12–14, 2018 at the Los Angeles Convention Center, with many companies holding press conferences in the days prior.
With the industry still in the middle of the eighth generation of video game consoles, no new hardware was introduced, and publishers and developers principally focused on new games to be released in 2018, 2019, and beyond. But this years E3 was special one for the tech giant Sony, with massive releases like Spiderman and The Last of us Part 2. One wonders after all these years how will the public reception of these games be.
No need to look any further, we have analysed the tweets tagged #PlaystationE3 and here are our findings!
Mining Data
We wrote a skcript ;) using twitter developer SDK and python. We extracted the favorite count, retweet count, the tweet text of the tweets having hashtag #PlayStationE3 between June 11 and June 13 2018. We ignored retweets and solely focused on the original tweets.
The script streams original tweets of people containing the hashtag and stores the tweets in a local file to be analyzed later.
We then cleaned the tweets off URLs and hashtags and mentions to make it data friendly. The analysis was also done in Python due to its ease of use and well supported data analytics libraries for data manipulation and visualization
Feature Engineering
The next step was to get the games that were mentioned in every tweet and create a new column in the dataset containing the same. This will allow us to analyze the data with respect to each game as well.
We used fuzzy algorithms to extract the names of the games that were mentioned in the tweet and added it as another feature.
Sentiment analysis was then performed on the tweets. We leveraged the Textblob library’s simple API to detect sentiment polarity for this. This feature was another important feature that can give interesting information regarding the event and people’s views towards the games. This feature was later analyzed extensively.
Results of the analysis
We observe that there are more tweets on 12th than the other dates. This was when the event took place.
Distribution of the sentiment polarity across the data:
We see that the majority of the people were almost neutral to positive in their tweets.
The image reveals some of the most commonly used words in the tweets.
People are more neutral during the release day but polarity is also more diverse on the same day.
Looks like people are more hyped about The Last Of Us Part 2, Spiderman and Hideo kojima’s Death Stranding. Surprisingly Sucker Punch Studios Ghost of Tushima doesn’t fall behind the hype train.
This figure shows the favorites and retweets per game, the hype trend continues with Spiderman, The Last of us part 2, and the memes on Death Stranding seem to be in full effect after Hideo Kojima was released from Konami entertainment.
A more in depth analysis of every tweet has been projected here, every dot here represents a tweet. We can clearly observe that majority of people were positive about Spiderman, Ghost Of Tushima and Control. Resident Evil 2 seems to be taking quite heavy criticism compared to the other games. The games on the hype train, Spiderman, The Last Of Us Part 2 (Ellie is all grown up!) and Death Stranding Have a little more variance as compared to the other games.
The violin plot here shows the spread and the frequency of the tweets sentiment.
Hopefully you enjoyed the results and found some of them interesting and some of them surprising. Happy gaming and Papa bless y’all!