Data science

Data science

• 1.
  Data Science -by Sujeet
• 2.
  Data Optimization Prediction Mathematics, Statistics,Algorithms Quantitative (Forecasting) Qualitative (Classification) Languages: R, SAS, Python etc.
• 3.
  Big Data Storage Problem,Processing Problem Optimization Prediction Mathematics, Statistics, Algorithms Quantitative (Forecasting) Qualitative (Classification) Languages: R, SAS, Python etc.
• 4.
  Data Optimization Prediction Mathematics, Statistics,Algorithms Quantitative (Forecasting) Qualitative (Classification) Hadoop Architecture, Map Reduce for processing Languages: R, SAS, Python etc.
• 5.
  Real Time Bidding -A Real Problem
• 6.
  Real Time Bidding DSP1 DSP 2 DSP N A B C D E F G H I SSP Advertisers Demand Side Platform Supply Side Platform Webpage Viewers Bid Request Bid Response
• 7.
  Request & Response Requset: userinformation as url For example : OS, Location, Gender, Age, Language, browser..etc. Response: Advertisement creative url, Bid Price
• 8.
  Advertisers Input Data •Target Segment • Budget • Flight Period(Campaign Period) • Optimization goals and directives
• 9.
  Advertiser’s Goal • MinimizeCost per Impression • Minimize Cost per Click • Minimize Cost per View
• 10.
  Advertiser’s Optimization Directives •Even Budget spent • Fast Budget spent • Aggressive Budget spent
• 11.
  Problem Statement What isthe number of bids an advertiser should put at certain price in a certain part of the day so that it achieves its goal while respecting its budget constraints.
• 12.
  Advertiser’s Expectation • Anadvertiser wants to know what is the average bidding price going on before bidding. • An advertiser wants to know which are the favourable parts of a day where he should bid.
• 13.
  Certain Definition • CPM(costper 1000 Impression) = Budget*1000/ Impression • CTR( Click Through Rate) = Clicks/Impression • Win rate = Number of Wins/ Number of Bids
• 14.
  Thank You