Initial reactions to the iPhone 4 have been pretty spectacular, with an estimated 1.5 million units sold during the first day andlines that lasted hours, but what was reception like on Twitter? While we know that the World Cup and President Obama topped the iPhone 4 pre-buzz, we still wanted to take a look at some deeper Twitter analysis.

Real-time social media tracker Trendrr was nice enough to offer us some data of the iPhone 4 and also AT&T over various periods of time.

Trendrr’s analysis covers both quantity and sentiment analysis, and we’ll examine those data points.

---

iPhone Tweets By the Numbers

---

Looking at the data on mentions of “iPhone” from November 2008 to the present, a few points really stood out. First, we should note that as of right now, Trendrr does not have full data for any days after June 22, 2010. We have hourly statistics (see below), but not full data info for an entire day.

iPhone (Matching Twitter Posts per Day)

Key: iphone (source: Matching Twitter Posts per Day)

>

Note the two gigantic spikes in the graph. The first occurs on June 17, 2009. This was the day that the iOS (or iPhone OS) 3.0 software was released. The next spike was on June 7, 2010, the official announcement of the iPhone 4. On June 7, Trendrr says that 653,000 tweets containing the phrase “iPhone” were sent.

Next, let’s take a look at an hour-by-hour examination, which started on Saturday and ran through the 25th.

iPhone (Matching Twitter Posts per Hour)

Key: iphone (source: Matching Twitter Posts per Hour)

Again, note the big spikes throughout the day of the 24th, with the phrase “iPhone” peaking at 55,537 tweets per hour.

---

AT&T and iPhone Sentiment Analysis

---

We find sentiment analysis really interesting, and the figures for the iPhone 4 offer some food for thought. While some have expected the overall net sentiment to be positive, the results are actually quite different.

iPhone (Twitter Sentiment)

Key: iphone (source: Twitter Sentiment % Negative) iphone (source: Twitter Sentiment % Positive)

Sentiment started out more positive than not for the term “iPhone,” but negative remarks started showing up on Monday, June 21. At that time, 25% of all tweets about the iPhone were deemed “negative.” However, this was before the iPhone 4 started to arrive to users, so we’re not really sure what contributed to the negativity, other than overexposure, perhaps.

The negative tweets died down again, and we started to see more positive tweets on June 22 and June 23, with positive and negative mentions in roughly equal amounts on Wednesday. Here’s where it gets interesting: 28% of the iPhone-related tweets on June 24, 2010 were negative. How many of these negative tweets were related to long lines or sold out stores, we don’t know.

As of today, June, 25, the positive tweets have edged past the negative tweets once again.

Moving onto everyone’s favorite whipping boy, AT&T. AT&T is not a popular company. In fact, it’s an oft-cited reason that people don’t get an iPhone. Still, looking at the data that goes back to May, it does look like the company had a few bright spots.

AT&T (Twitter Sentiment)

Key: at&t (source: Twitter Sentiment % Positive) at&t (source: Twitter Sentiment % Negative)

We’re not really sure why AT&T had such big spikes on May 22, 2010 and June 6, 2010, but those are the only two days the positives outweighed the negatives. Not surprisingly, on June 15, 2010, the day the iPhone became available for pre-order, AT&T had a 35% negativity percentage. Maybe it was the failed activation system.

The negativity cooled down a bit, but as of June 24, 2010, the negative tweets had reached 44%, according to Trendrr.

While we’ve heard in the past that Twitter can be a killer box office predictor, the same might not be true for social media and television ratings.

Social media monitoring company Viralheat spent a month tracking social media mentions — a million of them, in fact — right around the time that nine major TV shows were airing their season finales. The company looked at data from Twitter,Facebook, YouTube, Google Buzz and other sources, including the sentiment of all these mentions, comments and updates.

Then Viralheat compared the social media buzz with the Nielsen TV ratings for each show. Does chatter on the web equate to eyeballs on TV screens?

Not necessarily. Their research determined:

• Online buzz does not always correlate with ratings.
• Sentiment analysis of social media chatter casts those mentions in a new light.
• Even when ratings are lukewarm (as was the case for the Lost season finale) social media chatter can be overwhelming in volume. But a lot of those mentions online expressed a negative sentiment.
• Some shows, including Dancing With The Stars, saw good TV ratings but not many mentions online.

All in all, Viralheat found that Nielsen ratings combined with social media sentiment analysis provided the best snapshot of how a show’s season finale performed with viewers.

So, for you would-be social media gurus, next time you’re trying to measure a brand’s impact on the social web, remember to include some sentiment analysis along with your headcount, or you could be missing out on some critical information.

Here’s the infographic with more interesting factoids than you can shake a stick at; click to see the full version.