Threat Level: green Handler on Duty: Rob VandenBrink

SANS ISC InfoSec Handlers Diary Blog


Sign Up for Free!   Forgot Password?
Log In or Sign Up for Free!

How to Use Twitter for Information Mining

Published: 2009-01-21
Last Updated: 2009-01-22 14:45:41 UTC
by Lenny Zeltser (Version: 1)
0 comment(s)

Twitter is fun. It's also a powerful research tool. People increasingly use Twitter to share advice, opinions, news, moods, concerns, facts, rumors, and everything else imaginable. Much of that data is public and available for mining.

Here's how to use Twitter to gather useful information about topics, companies, and individuals. (Some call this practice "twittermining.") I'll cover native Twitter features, as well as third-party tools with catchy names, such as 5and2fish, Twitter Venn, TwitterFriends, PeopleBrowsr , Twitturly, Twitter Spectrum, and others.

Most of the techniques mentioned here don't require you to be a registered Twitter user. If you use Twitter, consider what data tidbits you release there, and whether you need to be more careful.

Twitter Essentials

If you're new to Twitter, watch Common Craft's Twitter in Plain English video. You may also like the Twittering Tips for Beginners article from the NY Times.

While some people tweet about mundane happenings like "Heading to work," many use Twitter for micro-blogging more meaningful updates. Also, inherent to Tweeter is its ability to encourage and capture conversations that span themes, time and geographies.

Google is great for sifting through standard web pages and official news. It's OK for searching blogs, though many prefer specialized services such as Technorati. None of these services match Twitter's access to an unfiltered, real-time perspective on what people are thinking and doing. (IMHO)

Research Emerging Topics

Twitter allows you to search what its users are saying right now, even when "traditional" data sources are of little help. For instance, when the US Airways #1549 plane crashed the other week, you could read about it immediately on Twitter before CNN had coverage of the event. You could observe the news spread like wildfire in Twitterspace.

Interested in following general information security topics? Search for "security information OR data". Want to track a worm? Search for its names: "Downadup OR Conficker". When researching a wave of Roundcube 0-day vulnerability scans for an earlier diary, I used Twitter's search to scope the problem.

By convention, Twitter users tometimes add topic metadata to their tweets by using a hashtag, a.k.a tracker keyword. If I were tweeting about malicious software, I might add "#malware" to my message. Hashtags help you track a particular topic by searching Twitter for the desired keyword, e.g. "#malware". You can also track hashtags via the Hashtags site.

Research Your Company and Competition

You can use Twitter to keep an eye on what people are saying about your organization, or to track your competitors. For instance, you may search Twitter's messages for the company's name, e.g. "SANS Institute". You may also want to search for the name of your industry or sector: "security AND training".

If the company you'd like to track has a Twitter account, you may want to follow it. Many major brands are active on Twitter in an official capacity. If they're not, you can look for company employees that tweet using personal accounts. (If interested in competitive intelligence gathering, see my earlier note about LinkedIn.)

Visualize Twitter Topic Data

Use Twitter Venn to understand relationships between concepts. It lets you enter 2 or 3 keywords, then searches Twitter for them, and draws overlapping circles that indicate tweets with those keywords. For example, using this tool to search for "coke,pepsi" shows what Twitter users are saying about each drink. Of particular interest is the area where the circles overlap, because those tweets mention both drinks in one message. For enothe, you can examine phishing trends by searching for "phishing,mail,phone".

Use Twitter StreamGraph to understand what words were associated with a particular keyword over time. For instance, searching for "security" can help you understand the concepts Twitter users tied to this term; clicking on each "wave" of the StreamGraph will show you the associated tweets. To understand the words associated to two different concepts, use Twitter Spectrum; for example, try searching for "unix,windows" to see what people are saying about each OS.

To visually search tweets within a geographic area, use 5and2fish, a mash-up between Twitter and Google. If you find this service too slow, use Twitter's advanced search to limit query by location. You can also use the TwitterLocal application to watch public geo-tagged tweets for a location.

Research Individuals

Use Twitter's search to locate all tweets that mention a person's name. Even if the person doesn't have a Twitter account, but is being discussed on Twitter, such a search will yield results. Twitter's advanced search also lets you locate messages from and/or to specific people; for this, you need to specify the individuals' Twitter usernames.

When you know the person's name, you can find his or her Twitter account from Twitter's Find People page, if the person's Twitter profile includes the name. Once you located the stream of the person you wish follow, click the Follow button on their Twitter page. (Here's mine.) Note that some people protect their updates to control who may see them.

To see which URLs a person shared on Twitter, use Twitturly. To search Twitter bios of the individuals who follow a particular person, use TweepSearch.

To track a person's feeds across multiple social networking sites, consider FriendFeed. It can collect data from Twitter, Facebook, LinkedIn, YouTube, and many others. You can search FriendFeed, similarly to Twitter. Some people who protect their updates on Twitter choose to expose them on FriendFeed.

Analyze and Visualize Twitter People Data

Several on-line tools can help you analyze and visualize public data about a Twitter user. For instance, TweetStats can show you when the person tweets and how often, which Tweeter clients he or she uses, and which words come up frequently in the tweets.

When profiling a Twitter user, examine the person's relationships: whom he follows and who follows him. Previously-mentioned Twitter Venn can help you understand the relationship between sent and received messages between Tweeter users if you use "from:" and "to:" tags in your search, such as "to:lennyzeltser,from:lennyzeltser".

To compare the words two Twitter users include in their messages, use the previously-mentioned Twitter Spectrum search, crafting the query in the form "from:user1,from:user2".

Another tool, TwitterFriends, can help you see who the person interacts with and how; it also lets you compare one Tweeter user's stats to another's.

TweetWheel helps you understand which of the Tweet user's followers know each other. It does this by creating a beautiful wheel that highlights the persons' Twitter ties to each other.

Track Data With Alerts and Dashboards

Twitter automatically generates an RSS feed for a particular user's update stream and for Twitter search results. If you'd rather receive email notifications, consider signing up for Twilert.

When following a lot of people or search results on Twitter, you can easily get overwhelmed even when using an RSS or an email reader. In this case, consider a popular Twitter client TweetDeck. It lets you create a dashboard that splits Twitter updates into separate column according to your criteria.

An up-and-coming competitor to TweetDeck is PeopleBrowsr, which offers similar features via a website, without requiring a local application. PeopleBrowsr also supports other social networks, such as Facebook and LinkedIn, in addition to Twitter. As of this writing, PeopleBrowsr is in alpha; it’s a bit rough around the edges, but looks very promising.

Be Mindful as You Research and Tweet

As you gather information on Twitter, be mindful of others attempting to manipulate you into arriving at their conclusions by feeding you misinformation. Cross-check data and understand its sources. For more on this, see Is Twitter A Market Manipulator's Dream on the TwiTip blog. If the topic of reputational attacks interests you, also look at the SpinHunters blog.

If using Twitter to share information and stay in touch with your friends, be mindful of how others might misuse what you reveal about yourself, others, or your company. In the words of Wired magazine's Steven Levy, "No matter how innocuous your individual tweets, the aggregate ends up being the foundation of a scary-deep self-portrait. It's like a psychographic version of strip poker--I'm disrobing, 140 characters at a time."

Liked this? Post it to Twitter!

-- Lenny

Lenny Zeltser - Security Consulting

You're welcome to follow Lenny on Twitter. You can also track new Internet Storm Center diaries by following ISC on Twitter.

Keywords:
0 comment(s)
Diary Archives