サバイバル分析によるユーザーの リテンション(継続率)分析

サバイバル分析によるユーザーの リテンション(継続率)分析

Transcript

1. αόΠόϧ෼ੳʹΑΔϢʔβʔͷ Ϧςϯγϣϯ(ܧଓ཰)෼ੳ 2017/05/20@TokyoR ҆ా༸հɹ@yasu_919 1
2. ࣗݾ঺հ 2 ҆ా༸հ Data Scientist @Exploratory ౦ژେֶͰػցֶशɾࣗવݴޠॲཧΛઐ߈ ओʹexploratoryͷRύοέʔδΛ։ൃΛओಋ email: yosuke@exploratory.io twitter: @yasu_919
3. Exploratory dplyr, tidyrͷؔ਺ΛϕʔεʹσʔλΛՃ޻ ࣗ෼ͰRͷؔ਺Λఆٛͯ͠࢖͏͜ͱ΋ग़དྷΔ %>% Ͱͭͳ͕Ε͍ͯ͘ εςοϓ 3 ΧϥϜ͔ΒॲཧΛݺͼ ग़ͤΔUI ؆୯Ͱ௚ײతͳՄࢹԽػೳ
4. 4 survival ύοέʔδ ཁ͢ΔʹαόΠόϧʢੜଘʣ ෼ੳΛ͢ΔͨΊͷΞϧΰϦ ζϜ͕ͦΖͬͨύοέʔδ Terry Therneau ࡞ऀ Contains the core survival analysis routines, including definition of Surv objects, Kaplan-Meier and Aalen- Johansen (multi-state) curves, Cox models, and parametric accelerated failure time models. આ໌
5. 5 ੜଘ෼ੳ ΋ͱ΋ͱ͸ɺੜ෺͕Ͳ Ε͘Β͍ͷ࣌ؒܦաͰ ࢮΜͰ͍͔͘ΛϞσϧԽ ͨ͠΋ͷ
6. 6 Ԡ༻ Πϕϯτ͕࣌ؒܦաʹΑͬͯͲͷఔ౓ੜ ͡Δ͔ͱ͍͏ͷΛ֬཰ϞσϧԽͨ͠΋ͷ Ͱ • ϢʔβʔͷΩϟϯηϧ཰ͷਪҠ • ैۀһͷ཭৬཰ͷਪҠ • ػցͷނো཰ͷਪҠ ͳͲͷ෯޿͍Ԡ༻ઌ͕͋Δ
7. 7 DEMO DEMO
8. 8 Boot Camp σʔλαΠΤϯεɾϒʔτΩϟϯϓ։࠵ ୈ1ճ: 6/17(౔), 18(೔) جૅฤ ୈ2ճ: 6/20(Ր), 21(ਫ) ि຤൛: 6/17(౔), 18(೔), 24(౔), 25(೔) جૅ+࣮ફฤ ฏ೔൛: 6/27(Ր), 28(ਫ), 29(໦) Exploratory͕ࣾɺγϦίϯόϨʔ ͷFacebookຊࣾɺUCόʔΫϨʔ ౳Ͱߦ͖ͬͯͨτϨʔχϯάɾϓ ϩάϥϜΛ೔ຊʹͯެ։ ಺༰ https://exploratory.io/training-jp ৄࡉɾ͓ਃࠐΈ