Big Data describes exponential growth and information availability in our digital world.
The 3 Vs
Big Data is sorted into two broad types:
- Structured Data is organized in a mechanical and manageable way (e.g. Nielsen’s panel data). It can be searched and integrated with relative ease.
- Unstructured Data is raw and unorganized. It includes data from natural language, images, and video that are more cumbersome and costly to analyze (e.g. sentiment of social media posts).
Big Data often presents challenges – the sheer volume makes discerning the signal from the noise challenging. Many big data sets, although census-style, are incomplete and not fully representative of the population.
How does it work at Nielsen?
For Nielsen, consumers watch and buy in an ever-growing number of ways. More Internet-connected devices mean more unstructured data. Big Data analytics seek to produce actionable insights from these complex “data exhausts,” or the information produced often inadvertently by greater interconnectivity.
Nielsen's “Big Data-style” products include Nielsen Digital Ad Ratings, Nielsen Twitter TV Ratings, Nielsen Buyer Insights, and Nielsen Catalina Solutions. Each solution measures consumer activity using census data from real transactions/interactions.
Examples of Big Data Processes are
- Data Extracts: The process of retrieving data from a data source (usually an unstructured data source) for further processing or storage.
- Data Harmonization: The process of providing data in a clear and comparable format that can be integrated with client data.