4.4.5 Data volume versus data quality

• Data volume

  The amount of data being processed
• Data quality

  The accuracy and trustworthiness of the data
• The results from any data analysis are only as good as the original data that is being processed
• Sampling

  • Improves the quality of the data at the expense of the volume
  • Researchers often select the best quality data from reliable sources
  • Researchers often exclude extreme values as they are unlikely to be representative
• The sheer volume of data involved in big data processing means there is high probability that at least some of the data is of poor quality
• Veracity

  Ensuring the correctness and trustworthiness of the data
• Other 'Vs' describing big data

  • Value
  • Variability
  • Visualisation
• The 'toos'

  • Too much data for traditional databases
  • Too complex for conventional categorisation
  • Too many updates to the data
• Use of big data in 2016 US presidential campaign

  • Large volumes of data on voters
  • Constant updates on voter opinions
  • Variety of data sources including social media and credit card purchases