Last Updated: October 12, 2015 2

Freebase – Social Database

Found In: Knowledge Sharing

While it is true that the Internet now has many billions of web pages, much of this is very hard to find and often it is a lot of noise because of how difficult it is to search effectively. For any and every search term entered into for now what is the best search engine, Google, there are still millions of pages listed in the search results. And it is a nonsense to think anyone is likely to go past the first three pages of search results looking for what they want to know about. And when they cannot, people tend to search again with a refined or new search term, hoping to get relevant results in the first few pages of search results, and often they have to repeat this process many times because of the failure of search engines to truly know what we are seeking, and the sheer volume of noise in those many billions of web pages. And while some people continue to work on ways to digitize all knowledge accumulated by humans throughout history, including Google’s project to scan tens of millions of books into the Internet so that they can all be searched, and while yet others are working to refine the algorithm to give us all a more accurate and relevant search tool kit, there is another team collaborating on pulling together databases of anything that could be useful into a wiki, so that information and data from many sources is all pulled into one central place, away from the noise of the general internet.

MetaWeb Launches Freebase

In 2007 a company called Metaweb made a public announcement about the launch of Freebase. They described Freebase as “an openly shared database of the world’s knowledge”. They went on to explain that they planned for Freebase to be a massive and collaboratively edited database of cross linked data sets. Freebase is a growing community of interested members with expertise in databases, and by uploading database information into a wiki environment; each data entity can grow an accumulated textual discussion about the accommodated dataset. The wiki styled data model they say is perfect for establishing the relationships between the various data sets which are called “entities”. Freebase is all about providing the interface suitable for users who are not programmers. That is, lay people with interests in the data or ‘meta-data’ can provide general information about it. These support people can categorize and connect different data entities into meaningful ways to understand the information.

Described by Tim O’Reilly, the idea is to provide for a bottom up approach to organizing information, using the power and wisdom of web 2.0 which facilitates user input and the more traditional and structured world of databases and the semantic web. Sources for the data are many and varied, ranging from ChefMoz, MusicBrainz and Wikipedia itself, Freebase also willing accepts contributions from its user base across the board, and Freebase now boasts of having populated it with more than twenty two million data entities. Metaweb uses its proprietary source code for Freebase and this is the engine of how searches are made to interrogate the entities for desired results. In particular, the Metaweb Query Language (MQL) is used to query the database and this is fundamental to the success of Freebase.

Freebase is About Free and Open Use of Data

Freebase is built upon a graphical database infrastructure, which was created in house. Metaweb is widely respected for its utilization of a graphics based modeling. This means that nodes and a set of links replace what have traditionally been tables to establish the relationships between nodes. As a non-hierarchical structure, Freebase is able to model much more complexity in the relationships. Individual elements are much less constrained this way than a conventional database structure of hierarchical tables. Also true is that at any time, later users can simply bolt on new nodes and create new relationships with new objects because of its graphical structure.

The Freebase API Helps Third Party Developers

Developers and users alike are provided with a suite of APIs (Application programmer Interfaces) which enable third party application development to put hooks into the Freebase application so as to export and be able to query the data entities and their relationships quickly and with ease. These APIs go to the heart of the collaborative community commitment of the team supporting Freebase and it in no small way explains the rapid growth and acceptance of freebase as the project for people interested in the collective wisdom of useful information all being brought into one central repository. The APIs do have a daily quota of one hundred thousand read calls a day against the Freebase data, and ten thousand write calls per day. The management of Freebase recommends that if developers need more than this, then they should consider doing a data dump from freebase and then doing further manipulations of data sets locally rather than through the API.

Freebase is all about open data systems and it is made up of a large community of liked minded people who love to contribute to open data systems. Working with sets of data may not be to everyone’s taste, but the growing accumulation of information, and the graphical way it is all laid out is truly a remarkable achievement. This especially true in the science and medical fields, where research data is so easily made available for others to develop and interrogate it further, progressing the accumulation of information for the betterment of all people, even those not directly involved or aware of Freebase’s existence. Then again, there does seem to be something for everyone, including the humble sports spectator; given that there are many large data set entities related to sport, and we all know how sports nuts love their statistics. Interestingly, the arts and entertainment section does seem to be one of the most popular categories with many project teams listed in the graphical summary of projects currently being worked on.