各种来源的一致� Source credibility appropriateness of methodology and source bias

各种来源的一致? source credibility

This preview shows page 5 - 8 out of 16 pages.

(各种来源的一致性) – Source credibility, appropriateness of methodology and source bias (来源的可信度, 适当的方法和来源偏差) Primary data: collection purpose: for the problem at hand. Collection process: very involved. Collection cost: high. Collection time: long. Secondary Data: collection purpose: for other problems. Collection process: rapid and easy. Collection cost: relatively low. Collection time: short. From secondary data to BIGdata • No consensus on its definition (对其定义没有一致一键) – Sheer volume of data – Improve business decision making How does big data differ from conventional data? (大数据和传统数据的区 别)
Image of page 5
Volume: 2.5 exabytes per day (2012) Velocity: real-time information processing – Variety: tracking, user generated, GPS • Both internal and external What can big data do for us ? • Cold hard facts 冷酷的事实 • Linking different sources of data 链接不同的数据来源 • Prediction 预测 • Real time prediction 实时预测 Volume-tracking using scanner data 使用扫描仪数据跟踪 Sales at electronic checkout counter, aggregate by brand, size and price (电子结 账柜台的销售,按品牌,规模和价格汇总) Tracked for a representative sample of national supermarkets, drug stores and mass-merchandisers (跟踪有代表性的国家超市,药店和大规模采购商) To track macro trends in spending (跟踪宏观支出趋势) A particular form of syndicated data - single-source data (联合数据的一种特殊形 - 单一来源数据) • Definition - a large secondary data base which is collected from a single source, i.e. a single group of respondents • Example - The Roy Morgan Research Centre interviews 50 000+ people annually on issues such as demographics, media habits, finance, tourism • Allows researchers to study the connection between different aspects of customers’ life Week 4 For the purpose of the unit, there are mainly two types of fixed-alternative questions: Categorical questions that classify people into groups Metric questions that quantify the strength of feelings, emotions, attitude and beliefs Main difference between metric vs. categorical data Categorical – can NOT add, subtract, or compute averages for (even if it is coded as numbers, such as in marital status on the previous page)
Image of page 6
Nominal scale- no rank-ordering between values Ordinal scale – rank ordering between values Metric – data that you can add, subtract, and compute averages for Interval scale – most measures of the strength of feelings, emotions, attitude and beliefs are based on an interval scale. Zero value does not mean “nothing”. Ratio scale- to provide exact quantity measures of tangible things. Zero value means nothing.
Image of page 7
Image of page 8

You've reached the end of your free preview.

Want to read all 16 pages?

  • Two '16
  • representative, secondary data

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture