Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
メルカリにおける分析環境整備の取り組み
Search
nagai shinya
August 19, 2020
8
7.7k
メルカリにおける分析環境整備の取り組み
以下のイベントの発表資料です。
https://forkwell.connpass.com/event/182769/
nagai shinya
August 19, 2020
Tweet
Share
More Decks by nagai shinya
See All by nagai shinya
Analytics Engineeringチームを立ち上げて学んだこと
__hiza__
4
1.7k
1日50万件貯まるクエリのログを活かして、SQLの生成に挑戦している話
__hiza__
7
1.6k
Analytics Engineeringチームの目標管理
__hiza__
62
36k
データ整備の優先順位付けに役立つテクニック
__hiza__
5
2.8k
データマネジメントがちょっと楽になるBigQuery監査ログの使い方
__hiza__
0
5.1k
レガシー化したdata pipelineの廃止
__hiza__
0
980
LookerのDashboardをより柔軟に作る
__hiza__
0
1.5k
Featured
See All Featured
How to train your dragon (web standard)
notwaldorf
89
5.8k
Reflections from 52 weeks, 52 projects
jeffersonlam
348
20k
RailsConf & Balkan Ruby 2019: The Past, Present, and Future of Rails at GitHub
eileencodes
132
33k
Chrome DevTools: State of the Union 2024 - Debugging React & Beyond
addyosmani
3
240
Responsive Adventures: Dirty Tricks From The Dark Corners of Front-End
smashingmag
251
21k
CoffeeScript is Beautiful & I Never Want to Write Plain JavaScript Again
sstephenson
160
15k
Bootstrapping a Software Product
garrettdimon
PRO
305
110k
What’s in a name? Adding method to the madness
productmarketing
PRO
22
3.2k
Optimizing for Happiness
mojombo
376
70k
How To Stay Up To Date on Web Technology
chriscoyier
790
250k
How to Create Impact in a Changing Tech Landscape [PerfNow 2023]
tammyeverts
49
2.2k
Sharpening the Axe: The Primacy of Toolmaking
bcantrill
38
1.9k
Transcript
1 ϝϧΧϦʹ͓͚ΔੳڥඋͷऔΓΈ גࣜձࣾϝϧΧϦ / JP Data Analyst ӬҪ৳
2 Introduction
3 ! ӬҪ ৳ ! גࣜձࣾϝϧΧϦ / JP ! Data
Analyst ◦ ੳڥͷඋͳͲΛ୲ ࣗݾհ
4 ! ݱঢ় ◦ ͳͥվળʹऔΓΉͷ͔? ! ͋Γ͍ͨ࢟ ◦ վળͷαΠΫϧΛճ͍ͨ͠ɻ !
औΓΈ ◦ ϨΨγʔͳσʔληοτΛഇࢭ͢Δɻ ◦ ͦͷͨΊʹɺۀͱKPIͱج൫ΛηοτͰߟ͑Δɻ ΞδΣϯμ : ϝϧΧϦʹ͓͚Δੳڥͷඋͷࣄྫ
5 ݱঢ় | ͳͥվળʹऔΓΉͷ͔?
6 ! ج൫ ◦ BigQuery + Looker ! ن ◦
ΫΤϦ࣮ߦϢʔβʔ 700ਓҎ্/݄ ◦ ࢀর͞Ε͍ͯΔςʔϒϧ 100Ҏ্/݄ ◦ Analyst, PdM, ML, CS, ͳͲ ϝϧΧϦʹ͓͚Δσʔλͷར༻ঢ়گ
7 ! ଟ͘ͷਓ͕σʔλΛۀʹ͍ͬͯΔɻ ! ੳڥͷඋ → ·ͩ·ͩෆेɻ ! σʔλج൫Λվળ͢Δࣄ͕ɺੜ࢈ੑͷվળʹܨ͕Δɻ ϑΣʔζతʹɺ্ཱͪ͛ɺීٴɺͱ͍͏ΑΓӡ༻ͷ՝͕େ͖͍ɻ
ͳͥվળʹऔΓΉͷ͔?
8 ͋Γ͍ͨ࢟ | վળͷαΠΫϧΛճ͍ͨ͠
9 ͚͍ͨ͜͞ͱ → ෛͷαΠΫϧ ᶃ վળʹऔΓͳ͍ ᶅ վળʹ͑Δ ͕࣌ؒݮΔ ᶄ
ݱঢ়ҡ͕࣋ ͡Θ͡ΘେมʹͳΔ
10 Γ͍ͨ͜ͱ → վળͷαΠΫϧΛճ͢ ᶃ վળʹऔΓΉ ᶄ ݱঢ়ҡ͕࣋ ͪΐͬͱָʹͳΔ ᶅ
վળʹ͑Δ ͕࣌ؒ૿͑Δ
11 Γ͍ͨ͜ͱ → վળͷαΠΫϧΛճ͢ ᶃ վળʹऔΓΉ ᶄ ݱঢ়ҡ͕࣋ ͪΐͬͱָʹͳΔ ᶅ
վળʹ͑Δ ͕࣌ؒ૿͑Δ ᶆ ेͳ༨༟͕Ͱ͖Ε ɹ ɹ ߈ΊͷվળͰ͖Δ
12 Γ͍ͨ͜ͱ → վળͷαΠΫϧΛճ͢ ᶃ վળʹऔΓΉ ᶄ ݱঢ়ҡ͕࣋ ͪΐͬͱָʹͳΔ ᶅ
վળʹ͑Δ ͕࣌ؒ૿͑Δ ᶆ ेͳ༨༟͕Ͱ͖Ε ɹ ɹ ߈ΊͷվળͰ͖Δ ࠓɺऔΓΜͰ͍Δͷ͜͜
13 औΓΈ | ϨΨγʔͳσʔληοτΛഇࢭ͢Δ
14 2ͭͷpipeline ݩςʔϒϧ ੳςʔϒϧ(৽) ੳςʔϒϧ(ϨΨγʔ) BigQuery Production Production͔ΒBigQueryʹσʔ λΛίϐʔ Production͔Βͷpipeline͕
2ܥ౷͋Δ ϝϯςφϯείετ͕͔͞Ήɻ ࣈ͕߹Θͳ͘ͳΔɻ
15 ৽pipelineͷยد ݩςʔϒϧ ੳςʔϒϧ(৽) ੳςʔϒϧ(ϨΨγʔ) BigQuery Production ňϨΨγʔpipelineΛഇࢭ ৽pipelineʹยد͍ͤͨ͠ʼn
͔͠͠ɺϨΨγʔ͔Βܭࢉ͍ͯ͠ ΔKPIΛ͍ͬͯΔνʔϜ͕͋Δɻ ϨΨγʔͳpipelineͷॲཧΛ৽ pipelineͰ࠶ݱ͠Α͏ͱ͍͕ͯͨ͠ ग़དྷͳ͔ͬͨɻ
16 ͳͥϨΨγʔͳpipelineͷґଘ͕ΊΒΕͳ͍ͷ͔? → ϨΨγʔͳpipeline͔Βܭࢉ͍ͯ͠ΔKPIΛۀʹͬͯ ͍Δ͔Β ৽pipelineͷยد
17 ࣮ / ۀཁ݅྆໘͔Βཁ݅Λཧ → ຊ࣭తʹ৽pipelineͷσʔλͰସͰ͖ͦ͏ͩͱ໌ ͳͥͦͷKPI͕ඞཁͳͷ͔? ۀཁ݅ͷཧ
18 ! ۀཁ݅ΛݩʹKPIͷఆٛΛݟ͠ → ఆ্ٛϨΨγʔͷґଘΛͳͤͨ͘ɻ KPIͷఆ͔ٛΒݟ͢
19 औΓΈͷϙΠϯτ ج൫ ࢦඪ ۀ ͚ͩ͜͜มߋ͢Δͷ͍͠ ηοτͰߟ͑Δ ج൫ͱۀ ηοτͰߟ͑Δ
20 औΓΈͷൣғΛߜΓࠐΉ ᶃΛࢀরͨ͠Ϣʔβʔ … 200ਓ/݄ ᶄΛࢀরͨ͠Ϣʔβʔ … 2ਓ/݄ ͬͯΔਓͷਓʹ100ഒҎ্ͷࠩ →
ॏཁͳͱ͜Ζ͔Βͬͨํ͕ྑ͍ table Tableผ ࢀরϢʔβʔ (ۙ30) ᶄ ᶃ
21 ·ͱΊ
22 ! ݱঢ় ◦ ͳͥվળʹऔΓΉͷ͔? → ੜ࢈ੑͷվળʹܨ͕Δ ! ͋Γ͍ͨ࢟ ◦
վળͷαΠΫϧΛճ͍ͨ͠ → ࣌ؒͷ࠶ࢿΛଓ͚Δ ! औΓΈ ◦ ϨΨγʔͳσʔληοτΛഇࢭ͢Δɻ ◦ ͦͷͨΊʹɺۀͱKPIͱج൫ΛηοτͰߟ͑Δɻ ·ͱΊ : ϝϧΧϦʹ͓͚Δੳڥͷඋͷࣄྫ