Case [practice] 11 double electricity supplier discount parsing routines

 

Analysis of target

  • Look at Taobao dual 11 businesses involved in how to warm, whether it is really a discount, how to discount the intensity

Ideas decomposition  

  • part1, double 11 businesses involved in how to warm? (Two-day sale + 11 pre-sale)
    • Look at the number and proportion of cases of goods sold in the two-day 11
    • Look at the brand number and proportion two-day goods sold in 11
    • Dual 11 day not in the sale of goods, how trends before and after? (Temporarily off the shelf? Presale? Or renaming shelves?)
    • Commodity true dual 11 activities to participate in, what brand? The number of items each brand real participation in the activities of 11 double distribution of what?
      • Note: participate in the activities of real commodities = double eleven day sale of goods + sale of goods (and then adding to the weight, remove the pre-sale and the sale of goods the same day)

 

  • part2, is it really a discount, how to discount efforts?
    • For each product, its price changes were observed before and after the double 11, double 11 to see whether its lowest price during the day, that is, whether the discount in double 11
    • For discounted merchandise, look at how many of its discount rate
    • And then rise to the brand (store) dimensions, look at its discount efforts to arrange what percentage of merchandise discounts, discounted merchandise and how these average discount rate

 

  • part3, summarizes the different brands (merchants) discount routines and strength, and greater efforts were extracted discounts and discount false list
    • Summary of different brands (merchants) and intensity routines discounts
    • Discounts are given greater efforts recommended product and brand list
    • Give false discounted merchandise and brand blacklist

analyzing tool

  • Python 3.7.3 (default, Mar 27 2019, 17:13:21) [MSC v.1915 64 bit (AMD64)]
  • Excel

 

the data shows

  • Two-eleven Beauty data from the 2016 Lynx analysis, raw data .xlsx format
  • Including update_time / id / title / price / store name, a total of five fields, where id is the unique identification of goods, its name is the brand name
  • This case mainly for training with the aim familiar with Python and related libraries, and to analyze the idea to practice this scene

 

Text as follows:

 

Guess you like

Origin www.cnblogs.com/zwt20120701/p/11284459.html