Google Refine 2.0 Aims to Clean Up Messy Data Sets

Google has launched Refine 2.0, a new project based on the existing Freebase Gridworks it acquired this summer along with Metaweb. The tool has been renamed Refine and version 2.0 has been released. With Google behind it, the tool is bound to be a lot more visible so many users may only be discovering it now, but there are new features for existing users as well.

"Our acquisition of Metaweb back in July also brought along Freebase Gridworks, an open source software project for cleaning and enhancing entire data sets," David Huynh, Google Search Infrastructure, said.

"Today we’re announcing that the project has been renamed to Google Refine and version 2.0 is now available," he announced.

"Version 2.0 introduces a new extensions architecture, a reconciliation framework for linking records to other databases (like Freebase), and a ton of new transformation commands and expressions," he explained.

Google Refine is a tool for those working with large data sets from various sources. Refine, as the new name implies, is useful for sifting through the data and cleaning it up, fixing inconsistencies and merging different data sets.

The previous Freebase Gridworks 1.0 has been appreciated by its users and came especially handy when doing research, in particular with public data from government agencies.

Refine 2.0 can fix issues like misspelled labels, different terms used for the same thing, making it easier to group relevant data.

It also provides ways of visualizing data to make it easier to analyze. You can check out the video below to see the tool in action. You can find the other two videos in the series here.

Just like Freebase Gridworks, Google Refine is offered for free as an open source project.

Google acquired Metaweb, the company behind Freebase and the Freebase Gridworks tool, last summer, interested in the value its data could add to things like structured data search.