When no other data cleaning tool could withstand the challenges from the world’s ugliest data sets, our engineers were inspired to build Transform. Transform takes in mountains of unstructured, mismatched data and converts it into something usable.
Transform uses machine learning to identify issues and suggest changes. This takes away much of the guesswork, manual effort, and time in data cleaning and structuring.
Designed by our data analysts to solve their toughest problems, Transform is built for scale. Transform can process, clean, and structure gigabytes of complex, unstructured data in no time.
Transform uses entity recognition to figure out whether a piece of text is a district name, school name, person's name, or country name, for example. With this knowledge, Transform can run complex fuzzy logic to match unique entities like geographies or individuals.
Transform can geocode and reverse geocode address data, identify geo-boundaries of geographic areas, or triangulate geo-locations from differing location APIs like Google Maps and HERE Maps.
It was an absolute pleasure to work with SocialCops on a project that was ambitious in scale — the extraction of information on about 12 million people from Delhi's voter rolls. The size of the project and the unique challenges of the data format would have made this project impossible for many others.
Our project had a demanding geospatial scope and SocialCops did a great job mapping data between census maps and the national databases. Using their platform, we were able to structure large volumes of data up to the sub-district level and visualize the data on an interactive dashboard.