Big Data – Its Relevance and Impact on Industrial Statistics (UNIDO Working Paper)

Several research projects have demonstrated the possibilities Big Data offer to official statistics (UNECE 2016). However, most of the research activities involve satellite images for geospatial use of statistics or mobile phone data to improve tourism and migration statistics. The use of Big Data in an economic sense has not yet been thoroughly investigated and is mostly represented by internet sales data on prices or aggregated private data due to proprietary and privacy concerns by the owners of Big Data. Thus applications of alternative data sources to industrial statistics, in particular to manufacturing-related activities, remain unexplored.

Although no concrete steps for the implementation of Big Data in industrial statistics have been taken so far, several options in this direction are possible. Incorporating Big Data or other alternative data sources into industrial statistics may entail similar problems as  described in this paper for the general use of Big Data in official statistics. Big Data are not a panacea for industrial statistics because they may only represent a source of data available for short-term estimates. Many applications would require an individual approach based on characteristic features of the target variable. Filling data gaps might be problematic since the missing values could appear in both data sources, in official ones and in alternative data sources.

Source: UNIDO Statistics Data Portal