The Value of Data Fusion

Recently, the PRISM team has been out on the road discussing our services and methods with potential and current clients. During the process, we generated interest in our data FUSION process and the value it adds to investigations. It made us realize our readers (you) could benefit from the same discussion. To simplify the process, we will discuss the value added from each type of data source. Currently, we utilize three main sources: Public Records, Online Content, and Social Media Content.

Public Records

Public records push a research project off in the right direction. With the abundance of personal information available, these records make it possible for you to look up criminal and litigation history, identify physical locations and property information, map out potential networks of family and associates, and vet out current licensures. Together, all of these records combine to build a base for your research. They provide relevant search terms and locations for the rest of your work. However, many public records databases contain outdated, incomplete, and inaccurate information. You still have to look at other sources of information to verify the data and paint a more complete picture of the individual of interest.

Online Content

General online content is a great way to find out how the individual of interest interacts with society at large. News stories can help research a multitude of issues related to an individual. First, they can help you identify any additional criminal or civil litigation information which may have been excluded from public records. Second, news articles provide a wealth of knowledge regarding business operations which can help you set a general context for the individual’s professional life. Finally, local news also tends to spend time giving lip service to those doing work in the community, which can help you identify additional organizations connected to the person of interest. When you combine this data with personal and professional websites, you have additional contextual data that informs you about what the person is involved in on a routine basis. However, this data, even when combined with public records, tells you very little about the actual person you are investigating and still may contain some inaccurate information.

Social Media Content

Social media content’s best value added comes from the data about the person’s personality, interests, and habits. When a person posts a wide array of information about the things they like, do, and consume, it becomes very easy to identify additional places to locate information. Social media profiles also provide a great wealth of connection data that allows you to confirm connections from public records and build out additional social networks. Additionally, you can verify some of the data that was originally identified in both public records and online content. It is imperative this data is fused together with public records and online data during the research process. Without social media, it is nearly impossible to fully understand an individual.

Conclusion

To add reliable, actionable intelligence to your investigation, it is imperative to use a data fusion approach. Without using a comprehensive online research solution, you can be acting upon inaccurate or incomplete information. In order to provide a truly thorough investigative service to your clients, you should implement a data fusion process as a integral part of your operations.

Building Your Tool Chest: Image Tools

Building Your Tool Chest is a series devoted to the review and analysis of tools that assist with social media and open source research and analytics. 

The world of social media and online content is becoming increasingly visual. Photos and videos are becoming the primary vehicle used by many to communicate their thoughts and ideas with others. Gone are the walls of text once prevalent in the era of Geocities. Today, online content generally focuses on conveying messages with little text in a manner designed to impact users in a big way.

This trend is exemplified by Facebook’s recent announcement it was transitioning to a more visual news feed. The change was reportedly prompted by the growing popularity of other social media sites such as Instagram, Pinterest, and Tumblr. Additionally, Google recently added a carousel design to its search returns which show photos related to search terms. These changes to a more visual culture has made the analysis of images during social media research much more important.

Analysts must be able to draw conclusions from the content of the photos themselves. More importantly, analysts must equip their toolkit with applications that visualize the metadata behind photos and give users the ability to search for images across the Internet. There are a variety of individual services that perform these functions such as TinEye and metapicz. However, there is a Google Chrome application which combines a variety of these image analysis tools into a single, simple interface: Image Tools.

Image Tools combines applications that extract exchangeable image file format (exif) data and provide reverse image search services together in a single right-click interface. With this application, analysts can gain insight into the data behind photos and determine if they are used elsewhere. The two main features of the tool are exif viewers and reverse image searches.

Exif Viewers

Exif data is a structured data set embedded into most digital images by the device that captured it. Certain tools can extract latent information from images for further analysis. To detail this concept, we will examine the exif data behind our Project Manager’s stylish sunglasses:

 Using Jeffrey’s Exif Viewer, you can visualize an abundance of information. You now know the photo was taken on an iPhone 3G running out of date software on June 19, 2013 at 12:56PM at a specific location. This information is vital when building timelines and placing people at certain locations. The exif tools featured in Image Tools also allow users to see a map of the area where the photo was taken. These latent details give analysts greater insight into the subject of their investigation.

 However, there are limits exif data. This data is best extracted from images downloaded from the Internet and email attachments. When an image has been uploaded to a site like Facebook or Imgur, the rich exif metadata is generally removed from the photo. Ultimately, the ability to extract this data depends on the website hosting the image and whether users or the site have scrubbed the metadata from the photo.

Reverse Image Searches

Reverse image searches are valuable when you are attempting to identify online accounts maintained by a single individual or the original source of an image. The toolset available in Image Tools allows you to right click on an image and search for sites where that image was used before. This saves a great deal of effort when trying to build out an individual’s online footprint.

To show how this works, we will use a reverse image search to identify accounts belonging to a member of the Syrian Electronic Army. Once you identify one account maintained by the person of interest, you can right click on their profile image and perform a reverse image search. The search for the image below returned thirteen results.

From the results, you can to identify YouTube and LinkedIn accounts. Image Tools gives you a jump-start on building out this individual’s digital footprint by helping you identify additional accounts within minutes. From there, you can continue to build out this individual’s network based upon the other data found on these profiles.

Without facial recognition capabilities, the average profile photo does not fare so well in a reverse image search. This method works best when the individual’s avatar is somewhat unique. However, you should always run a reverse image search because there may be relevant results. Using these searches can be valuable time savers when trying to identify where avatars may have been used elsewhere.

Conclusion

The two sets of features detailed above show how analysts can exploit the growing image-centric nature of online content when conducting their research. Image Tools is a great all-in-one app for viewing exif data and searching through disparate data sources to locate images. Every analyst’s toolkit should have a similar tool to investigate the data behind digital images. Thanks to exif data and reverse image searches, you can go beyond what it is in a photo by examining what is behind it as well.

Building Your Tool Chest: The New Google Maps

Building Your Tool Chest is a series devoted to the review and analysis of tools that assist with social media and open source research and analytics. 

In May, Google revealed the new Google Maps interface. Currently, it is in its testing phase, and no firm release data has been set. A few of us in our St. Augustine office started using it and instantly noticed graphical improvements to the system. There are some fantastic new visualization features, including tilt view and additional photo streams. However, some ease of use issues may make this version more cumbersome for research. We used Flagler College as a comparison point for differences between the new and current Google Maps.

Viewing Locations is a Snap

Now that the giant sidebar is gone, it’s simply easier to view the map. The integration of the photo carousel allows you to quickly view the physical appearance of the location of interest. The compass and tilt view are great for changing the perspective of the map to fit your needs. The tilted and rotated screenshot above features Flagler College in the lower-center of the map while looking east towards Anastasia Island and the Atlantic Ocean. Although the 3D graphics do not always properly render, it can give you a good perspective if you’re trying to learn the lay of the land.

Getting Location Data is More Difficult

 

In the current version of Google Maps, a static bubble pops out of the pin when you click on a location of interest. It has a clearly defined place to click for “more info,” which takes you to the location’s Google Plus page. However, in the new Google Maps, it is not that simple. The information appears in the upper left hand corner of the page in a collapsing drop down. In order to see this information, you have to keep your cursor over it. In addition, the clearly defined “more info” link is missing; instead, the user must click on the reviews to be taken to Google Plus.

Location Mapping in Earth View is a Bit Wonky

As recently mentioned in Slate, Google’s idea to plot locations on maps based upon users’ preferences and history could prove to be problematic, especially for researchers. Personalized data can exclude items which may not seem relevant to our personal interests even though we may need a complete data set for area mapping and analytics.

In map view, both the new and current Google Maps show comparable sized data sets and matching data points. However, when we move into Earth view in the new version and satellite view in the current version, we begin to see differences in data. The current satellite view provides more data points on the map, while the new Earth view shows only selective locations. It is hard to tell if this is a symptom of personalized data or a glitch in Earth view.

Conclusion

Without knowing all of the changes Google will make to the new maps system before it is released, it is hard to know how many of these functionality issues may impact our work in the future. Additionally, it is difficult to predict future changes in Google Maps based upon changes in the business environment, as social companies are constantly outmaneuvering one another through acquisition and development. Google’s recent acquisition of Waze could further impact direction and travel features added to new Google Maps, which is why we did not include those features in our discussion. Between the acquisition and feedback being provided to Google during the testing phase, we could see many changes to new Google Maps before it is finalized.

Although we will not know the full impact of the changes to Google Maps until it is released, accessing tools during the testing phase keeps us ahead of the curve by anticipating how changes in technology may impact our future research practices. By providing feedback during the preview process, it allows us to shape the development of these tools for easier and expanded use by analysts. If you’re interested in testing out the new Google Maps, click here for an invite and more information about the tool.

Tips & Tricks: 4 Faster Ways to Identify Data Sources

Tips & Tricks is a series of helpful techniques designed to increase efficiency in open source research.

During investigations, it is pertinent that we work smarter, not harder. One of the tricks we have learned is to find sources which aggregate social media profile and website information. It decreases the time spent identifying individual accounts and allows us to devote more time to data analysis. There are four major sources we use: websites & blogs, social media profiles, social influence sites, and checkusernames.com. To illustrate how it works, we chose to research Britney Spears.

Websites & Blogs

Many individuals and organizations now maintain their own website and/or blog, even if it is something as simple as an about.me page. On these sites, individuals tend to identify social media profiles and other additional sources of information. In the case of Britney Spears, the header of her website links to her Twitter, Facebook, YouTube, Google+, Tumblr, and iTunes profiles, and slightly down the page is a link to her Instagram profile. The front page of the website clearly identifies seven additional data sources.

Social Media Profiles

Today, most Americans maintain at least one social media account. However, many people maintain more than one account. With the advent of apps, there is a wealth of data from linked accounts and cross-posted application data. In the case of Britney’s Tumblr, six data sources were clearly labeled: her website, YouTube, Twitter, SoundCloud, Google+, and Facebook profiles.

If none of these accounts are clearly labeled, you also can identify profiles in two easy ways. First, you can look at the listed connected accounts on the profile, such as a Google+ profile listed on a YouTube channel. Second, you can identify profiles from posts pushed from another data source, such as an Instagram photo pushed to a Facebook feed.

Social Influence Sites

An often-overlooked source of information is social influence sites, such as Klout and PeerIndex. These sites allow users to measure their influence by connecting social media profiles to their account and gain perks for topic and network influence. As researchers, social influence sites allow us to quickly harvest links, establish connections between users, and identify topics of interest. 

On Britney’s Klout page, there are six linked social media accounts: Twitter, Facebook, Google+, YouTube, Instagram, and Tumblr.

Checkusernames.com

Once you have identified usernames for an individual, a quick and simple way to see if it is used on another network is to run it through checkusernames.com. If the site identifies the username as registered, it is possible the same user has a profile on an additional site. However, as some usernames are popular, it is important to thoroughly vet the profiles to ensure the same user maintains the accounts.

Conclusion

While Britney Spears is more plugged in than the average individual, the methodology still remains true. Capturing connections between data sources fundamentally expedites the identification process. Because there is no master tool which identifies profiles and websites, using disparate data sources to capture this information is the key to building and analyzing an individual’s complete digital footprint.