How Does It Work?

The Mannheim Search Joins Engine is a search engine for tabular data. It takes list of keys and optionally a property name as input and then extends it with information from various sources on the Web. To accomplish this, we are using the biggest non-commercial corpus of Web tables that is currently available, large corpus of Web tables extracted from Wikipedia, and two large linked dataset extracted from the whole Web. In addition to automating the task of extension, the Mannheim Search Joins engine also takes care of data quality assessment and data integration. Simply trusting the first search result is obviously not a safe way to get high quality data. Hence, the Search Joins engine combines the values of thousands of tables and applies state-of-the art data fusion techniques. This is also supported by our built-in data integration features, which for example transform values in different units (like 'miles' and 'kilometers') and make them comparable. Finally, the Mannheim Search Joins engine provides end-to-end provenance tracking, meaning that the user always knows where the data come from.