Gio Wiederhold PDM 4
1. Data Collection
•Two choices
1.(rare) Collect data specifically for analysis
§allows careful design --
Ømodel causes and effects
ØPurchase = f(price, color, size, custumer inc., gender,. ,,
×costly
×often small to make collection manageable
§imposes delays
2.(common) Use data collected for other purposes
§take advantage of what is readily available
Ølow cost
×filtering, reformatting, integration
§incomplete - rarely covers all causes / effects
§biased -- missing categories
×only people with phones, cars -- shopping in super markets