A Data Science Central Community
This blog was originally published on our Text Analysis blog, the blog post set out to analyze and visualize 11 million tweets collected around the time of and during Apple Live 2014.
Apple Live probably got off to the worst start possible earlier this year. Most of us who tried to log on to watch the much-anticipated launch were first, forced to watch the live feed in Safari and second, greeted with the TV Truck Schedule Screen...
To add to this Apple also made a complete mess of the audio. We were left sitting refreshing the page, waiting for the stream to start while being subjected to an audio visual nightmare, described brilliantly by this “fan” below:
To simulate the #applelive experience, open up several separate YouTube vids, play them simultaneously, minimize, stare at a test pattern.— Mark Reed-Edwards (@mreededwards)September 9, 2014
At AYLIEN, we gathered 11 million+ tweets mentioning 'Apple', 'iPhone', 'iOS', 'iPad', 'Mac', 'iPod', 'Macbook', 'iCloud', 'OS X', 'iWatch' and '#AppleLive' from the 4th of September to the 10th of September with a view of analyzing the tweets to gain insight into the voice of Apple Followers.
The Tweets over time graph covers tweets from 4pm on launch day to 12am that night. The top half of the graph displays the volume of tweets by journalists and the bottom displays tweets by the general public.
What’s interesting here is the difference in what the two groups got excited and most vocal about. The general public was most vocal on Twitter in anticipation of the iPhone updates, not the Keynote, not the Apple Watch, not Apple Pay or not even one of the biggest rock bands in the world announcing a free album giveaway to over 500 million people, no, just an announcement to say the iPhone is going to be bigger! The journalists however, as a group were far less biased towards any one announcement and tweeted about what they thought were the biggest announcements of the day and not surprisingly, given recent hype in the tech industry around payments, they were most vocal about Apple Pay.
The disappointment of #AppleLive was evident on Twitter especially at the start of the launch.
Running Sentiment Analysis on the tweets we can see from the graph above that the overall polarity of tweets took a sharp dive into the negative in the build up to the event. The public weren’t afraid of expressing their opinion as to how frustrated they were to be missing out on the action and Apple’s Audiovisual team must have been scrambling to make things right.
The graph above shows the polarity of tweets throughout the launch. It’s pretty clear the difficulty people had streaming the live cast put a dampener on the mood overall with the sharp switch from positive to negative around 5 pm. However, Apple did manage to turn things round with some well received announcements which brought them back into the green. While it’s interesting to see the overall polarity of the tweets, things start to get more interesting when we look at what aspects in particular about the event were p*ssing people off?
Whether you are analyzing Tweets, Articles or Reviews the overall sentiment and knowing whether it is positive, negative or neutral is cool and useful to know. However, Sentiment Analysis gets a lot more valuable and interesting when we can identify what aspects of entities in particular are positive or negative. Aspect Level Sentiment Analysis refers to identifying opinions or sentiments expressed on different features or aspects of entities; a phone, a camera, a bank. The aspects or features of entities would be say, the screen of a phone or the battery of a watch.
Consider the following tweet as an example: “The iPhone 5s is amazing, I love the camera and the OS is so fast. The battery is terrible though.”
The entity here is the iPhone and the aspects are the Camera, OS and Battery. If you were to focus on tweet-level sentiment for this particular example it would most likely be tagged as a positive tweet and you would most probably miss out on the negativity related to the battery.
The word cloud below highlights what in particular were the negative aspects of the Tweets about the launch. This information would be important to Apple in analyzing their brand, the overall success of the event and to get an understanding of what in particular people liked and disliked about the event. The same could also be done for particular products that were announced.
For the purpose of this blog we decided to focus on one aspect of the launch in particular from the Apple Live event: the Apple Watch.
The word clouds relating to the Apple watch highlight some key points which give a strong insight into the public reaction to the Apple Watch.
In saying that it’s also pretty interesting that both the positive and negative tweets tell us that most people just care about how it looks over any of its fancy features!
Sentiment Analysis doesn’t have to stop at high level understanding of whether or not a sample of tweets, reviews, emails, NPS scores etc are positive or negative. The ability to dive deeper into the “what” and “why” of positive or negative sentiment allows us to gain a better understanding of public opinion.
Whether we are looking for indications of what features people like and dislike in a product or offering or what aspects they place most importance on we can gain a deeper insight into opinions by focusing on aspects in Sentiment Analysis.
About AYLIEN: We are a Text Analysis company who have built a Text Analysis API, among other products, designed to help developers, data scientists, business people and academics extract meaning from text. You can try out our API here.