Social Media

Facebook Graph Search: History, Function & OSINT Use

Understand how Facebook Graph Search uses semantic processing to query the social graph. Explore its history, OSINT applications, and manual usage.

1.3k
facebook graph search
Monthly Search Volume

Facebook Graph Search was a semantic search engine designed to return answers to natural language queries instead of a traditional list of links. It allowed users to search the "social graph" to find specific people, places, and interests based on the vast data shared within the Facebook network.

Introduced in March 2013, the tool combined the data of over one billion users and external information to provide personalized search results. Unlike traditional keyword-based engines, it was optimized to understand phrases and relationships.

While the feature launched in Beta in January 2013 to a limited group of users, it eventually became a primary way for users to discover content within their friends' networks. However, Facebook began obscuring the tool's public visibility in late 2014 and deprecated the majority of its functionality in June 2019.

Why Facebook Graph Search matters

For marketers and researchers, the tool provided deep insights into audience behavior and connections.

  • Audience Research: Marketers used the search to find pages liked by specific demographics, such as "pages liked by people who like SEO."
  • Recruitment: Recruiters used queries to find people with specific job titles at certain companies (e.g., "NASA employees who are friends with Facebook employees").
  • Investigative Intelligence: Journalists and investigators used the tool to find photos and posts from specific locations or timeframes to verify events.
  • Network Mapping: Users could visualize the relationships between people, business pages, and geographic locations.

How Facebook Graph Search works

The mechanism relies on a semantic search algorithm that processes intent rather than just matching keywords. It searches for specific "objects" on the site, such as people, photos, or check-ins, and the connections between them.

  1. Semantic Processing: The engine analyzes natural language phrases to understand what the user wants to find.
  2. Social Context: Results are prioritized based on the user's network of friends and their privacy settings.
  3. Data Integration: The tool originally used specialized indexing to handle approximately 700 TB of post and comment data.
  4. External Pairing: For results outside of the social graph, Facebook partnered with Microsoft's Bing search engine from 2008 to 2014.

After the 2019 deprecation, users had to resort to manual URL construction. This involves creating a base URL (e.g., facebook.com/search/posts/?q=keyword&filters=) and appending a Base64-encoded JSON string to define specific filters like "Most Recent."

Best practices

Use English (US) settings. The search algorithm was originally built and optimized for the U.S. English version of Facebook. Switching your language settings ensures the most consistent results when using the native search bar or manual URL strings.

Locate specific ID numbers. To search for content tied to a specific user, page, or group, you must find their unique Facebook ID. This is typically done by viewing the page source and searching for terms like userID, pageID, or groupID.

Utilize standard URL parameters. If the UI does not show the filters you need, you can manually adjust the URL path. Common parameters include /search/people/ for user queries and /search/photos/ for image results.

Refine by location. You can limit results to specific coordinates or distances. This is useful for finding content or people tagged at a specific venue or within a city.

Common mistakes

Mistake: Trying to use Graph Search on a mobile device. Fix: Use a desktop or laptop, as the advanced search features and manual URL constructions are not available on mobile interfaces.

Mistake: Searching for private information. Fix: Respect privacy constraints. The tool only surfaces information that is already public or shared with the searching user; it does not bypass a user's chosen privacy settings.

Mistake: Using "curly" or tilted quotation marks in JSON strings. Fix: Always use straight quotation marks when building manual search filters to avoid "Page Not Found" errors.

Mistake: Expecting keyword matches for phrases. Fix: Remember that this is a semantic engine. If a direct keyword search fails, try a natural language phrase like "Photos of friends in Paris."

Examples

  • Example Scenario (Marketing Research): A strategist wants to find which interests are common among fans of a brand. They use the query "Pages liked by people who like [Brand Name]" to identify potential partnership opportunities.
  • Example Scenario (Recruiting): A recruiter searches for "People who work at [Competitor Company] and live in Seattle" to build a list of local leads.
  • Example Scenario (OSINT): An investigator searches for "Photos taken at [Location Name] in June 2019" to find crowdsourced imagery of a specific public event.

FAQ

Why did Facebook shut down Graph Search? Facebook stated that the majority of users prefer keyword-based search over semantic phrases. There was also significant speculation in the tech community that the tool was paused or deprecated due to growing concerns regarding user privacy and data security.

Can I still search for "Photos liked by [User]"? While the original easy-to-use search bar shortcuts were removed, many of these queries can still be executed through manual URL construction using the user's ID and specific Base64-encoded filters. Tools like Graph.tips were developed to help automate this process for the investigative community.

How do I find a Facebook User ID? You can find an ID by right-clicking on a profile, selecting "View Page Source," and using Ctrl+F to search for entity_id or userID. The number following that tag is the unique identifier needed for advanced queries.

What happened to the Bing partnership? Facebook ended its deal with Bing in December 2014. The platform moved away from showing external web results within its social search interface to focus more on its internal keyword search improvements.

Does Graph Search work in languages other than English? The corpus indicates that the original feature was restricted to the U.S. English version of Facebook. Users in other regions often had to change their account language settings to access the tool.

Start Your SEO Research in Seconds

5 free searches/day • No credit card needed • Access all features