Who cares about Bitcoin, look at Big data millions Twitter analysis

Derrick Harris analyzed a total of about 1.3 million tweets this February about Bitcoin or its founder, Satoshi Nakamoto, who wanted to learn from the information about people who were really interested in bitcoin and how Twitter's activities tracked big events. In addition, a detailed analysis of the publisher, post and specific information. This article comes from Gigaom, below look at Derrick Harris to bring us the wonderful explanation.

The following is the translation:

Summary: In order to understand the global phenomenon of Bitcoin, I analysed a total of about 1.3 million tweets about Bitcoin in February. The following is an analysis of tweets, posts, and specific information.

I recently analyzed a total of about 1.3 million tweets in February this year about Bitcoin or its founder Satoshi Nakamoto. I would like to learn from this information about people who are really interested in bitcoin (at least they are interested enough to push them to send it) and how Twitter's activities track big events.

Here's the information I got. Click on the picture to read more details.

(Disclaimer: I am neither a statistician nor a programmer, so I use simpler statistical tools and work with companies in other fields.) gnip--is now part of Twitter-providing this data. I use the cloud-based analytics services provided by Chartio for most quantitative and partial visualization. I through Google Fusion tables to achieve Marc Andreessen map, through the Tableau news content analysis, through the Alchemy API analysis Mt.Gox microblogging popularity.

Who's talking about Bitcoin on Twitter?

A total of 333,144 Twitter accounts have posted tweets about Bitcoin. But the number of tweets published by each account varies greatly. Some accounts are apparently junk-mail accounts or other types of robotic accounts-thousands of accounts have just posted a tweet or have not published tweets at all-and the rest are constantly focusing on bit rates and bits. The figure below is the most active 10 accounts in February.

The following figure is the maximum number of 10 users. CoinDesk, a journalist with no surprises, has topped the list. In fact, most of the numbers are higher (some even higher) because tweets often refer to multiple Twitter accounts, but because of the format of the data, it's hard to compute those tweets.

Some accounts are just "in the Pan", and a tweet has suddenly been massively forwarded, such as talk show host Conan O ' Brien.

This tweet with a user named Bacon Bangkok.

Marc Andreessen is the most active and highly watched

I think it is also necessary to analyze the situation of the plus V users to see which celebrities are discussing bitcoin on Twitter. The next figure is the most active plus v user in February, with VC Marc Andreessen (@pmarca) topping the list.

Andreessen has also been mentioned many times, which is expected-he has been mentioned more than 3,600 times (including by other users)-considering that he has only released 145 tweets, a huge number. We show the huge network he created through an interactive table (click here). Small yellow dot representative mentions Andreessen and other people's tweets, the blue dot represents the Twitter account that publishes tweets. Andreessen is a rhubarb dot, he connects the people he mentions and the people he mentions.

You can see that the number of network users that Andreessen directly reply or push directly is very small.

where Bitcoin (possibly) prevails

The following figure shows the time zone with the largest number of tweets published, which is an approximate result, given the existence of a spam account and the absence of an explicit time zone or time zone error for some accounts. In addition, this statistic is for a single tweet, not for different Twitter accounts in different time zones. For example, 20% of the 15,000 tweets released by the ALLTHINGSBTC account are from the London time zone.

The problem can be very different if analyzed by user-defined locations. For example, you will find that there are thousands of statements to refer to New York, and Cryptogeeks represents #bitcoin #Litecoin #Altcoins.

If you want to see the visual effects of the data, the image below is a map of the small sample user location mapped using Google Fusion tables. This is not exactly the exact--fusion the tables view places everything, even if some have no real place--but it roughly describes the status of the global Bitcoin.

Mt.Gox collapse

The above chart describes who is publishing the tweets, but in order to fully analyze the form of the February Bitcoin, we also need to analyze the demise of Bitcoin trading service Mt.Gox.

Mt.Gox's death throes see the timeline below. February 7 is a plunge, the day Mt.Gox official announced a moratorium on Bitcoin withdrawals. February 10, Mt.Gox extended the ban on withdrawals, bitcoin prices plummeted. February 24 Night, Mt.Gox has ' lost ' 750,00 bitcoin, the market value of 375 million U.S. dollars, the other morning, the number of tweets soared. February 28, Mt.Gox filed for bankruptcy protection.

The following figure is analyzed in hours. Note that the number of spam accounts has soared since February 8.

Mt.Gox's announcement late on February 6, when the media began discussing the collapse of Mt.Gox, Twitter's complaints about withdrawals and claims that Mt.Gox was about to perish were growing. February 5 Mt.Gox on the devaluation of about 100 dollars is also one of the reasons.

February 4, the Swedish Pirate Party founder Rick Falkvinge claimed that Mt.vox has accumulated up to 38 million of the outstanding withdrawals (that is, Bitcoin left the user account but not paid to the user).

In short, the comments about Mt.Gox follow the latest news. In the chart below, the blue line represents negative tweets, yellow lines represent positive tweets, and purple lines represent all the tweets of the day about Bitcoin. By February 25, the number of negative tweets referring to Mt.Gox accounted for One-fourth of the number of Bitcoin tweets.

Count the reporters in.

When a journalist understands a story, their articles tend to spread widely. February, a total of 247,000 different connections were shared, shared 1.02 million times. I analyzed the top 10,705 connections that were shared most, and these links were shared at least 12 times. Of these, 1328 (less than 1% of the total number of links)--extent, formality--comes from 25 technical and general news sites. They were shared 121.931 times and accounted for 12% of the total number of analytical activities in February.

(A note on links: Given that rss, Google, or other social media will modify links when sharing links, it is difficult to get the exact number for each link.) I shortened cells to 100 features in Excel, which is a good way to deal with multiple linked cells. ）

To make a long story short: When assessing the popularity of a publication, using total sharing is more accurate than using a single link. Using this type of analysis for the first 44,810 links, we'll get multiple links for each publication, with a minimum of higher sharing.

Now let's see what we get when we join CoinDesk, the news site that focuses on the currency. Click here to view an interactive version of this chart.

Similarly, news sites do not match spam, Bitcoin wallet links, exchanges, monitors, or other non-news sites. The top 15 most frequently shared links are just such sites, totaling 76,534 times. You can click here to view. This article from Wired (which seems to replace the earlier version) is the most frequently shared link that has been shared 1404 times,

Salute to Satoshi.

Through the analysis of the tweets, I got the above information, the full text at the end.