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
AWS Loft Tokyo のASK AN EXPERT ブースにおけるご相談・ご対応ロ...
Search
Sponsored
·
Ship Features Fearlessly
Turn features on and off without deploys. Used by thousands of Ruby developers.
→
Eiji Shinohara
February 23, 2019
Technology
0
670
AWS Loft Tokyo の ASK AN EXPERT ブースにおける ご相談・ご対応ログ を分析しました :) / ASK AN EXPERT at AWS Loft Tokyo - Tech Consulting Log Analysis
delivered this talk at JAWS DAYS 2019 (
https://jawsdays2019.jaws-ug.jp/
)
Eiji Shinohara
February 23, 2019
Tweet
Share
More Decks by Eiji Shinohara
See All by Eiji Shinohara
Algolia Best Practices Fall 2020
shinodogg
2
1.2k
Algolia Fall 20 Release - wrap up in Japanese
shinodogg
0
950
Algolia 2020 Autumn
shinodogg
0
3.3k
Algolia introduction - DEMO and Ranking Formula
shinodogg
0
480
Introducing Algolia with Demo
shinodogg
0
6.1k
Algolia Announces Global Expansion Into Japan
shinodogg
0
2.5k
Introducing Algolia in a nutshell
shinodogg
1
1.3k
Building and Running Microservices with AWS
shinodogg
0
740
Accelerating AdTech on AWS in Japan
shinodogg
1
350
Other Decks in Technology
See All in Technology
SREのプラクティスを用いた3領域同時 マネジメントへの挑戦 〜SRE・情シス・セキュリティを統合した チーム運営術〜
coconala_engineer
2
770
ClickHouseはどのように大規模データを活用したAIエージェントを全社展開しているのか
mikimatsumoto
0
270
SRE Enabling戦記 - 急成長する組織にSREを浸透させる戦いの歴史
markie1009
0
170
AI駆動開発を事業のコアに置く
tasukuonizawa
1
360
ランサムウェア対策としてのpnpm導入のススメ
ishikawa_satoru
0
220
StrandsとNeptuneを使ってナレッジグラフを構築する
yakumo
1
120
SREじゃなかった僕らがenablingを通じて「SRE実践者」になるまでのリアル / SRE Kaigi 2026
aeonpeople
6
2.5k
[CV勉強会@関東 World Model 読み会] Orbis: Overcoming Challenges of Long-Horizon Prediction in Driving World Models (Mousakhan+, NeurIPS 2025)
abemii
0
150
30万人の同時アクセスに耐えたい!新サービスの盤石なリリースを支える負荷試験 / SRE Kaigi 2026
genda
4
1.4k
ECS障害を例に学ぶ、インシデント対応に備えたAIエージェントの育て方 / How to develop AI agents for incident response with ECS outage
iselegant
3
290
AIが実装する時代、人間は仕様と検証を設計する
gotalab555
1
180
Bill One急成長の舞台裏 開発組織が直面した失敗と教訓
sansantech
PRO
2
400
Featured
See All Featured
The SEO Collaboration Effect
kristinabergwall1
0
350
The Pragmatic Product Professional
lauravandoore
37
7.1k
Navigating the moral maze — ethical principles for Al-driven product design
skipperchong
2
250
Building Flexible Design Systems
yeseniaperezcruz
330
40k
Marketing to machines
jonoalderson
1
4.6k
Data-driven link building: lessons from a $708K investment (BrightonSEO talk)
szymonslowik
1
920
[RailsConf 2023] Rails as a piece of cake
palkan
59
6.3k
Future Trends and Review - Lecture 12 - Web Technologies (1019888BNR)
signer
PRO
0
3.2k
Un-Boring Meetings
codingconduct
0
200
The MySQL Ecosystem @ GitHub 2015
samlambert
251
13k
We Have a Design System, Now What?
morganepeng
54
8k
Mind Mapping
helmedeiros
PRO
0
90
Transcript
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. ) 9C :KN : E / L A
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. ( ) 2 7B : @ @ @ : A : @ @ @ 31 .4A4 T c 0 - a WS ML 31 /@8 2@ @ JI L J
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. AWS Loft Tokyo? ASK AN EXPERT?
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. ASK AN EXPERT @ AWS Loft Tokyo ! 2 01 8 - AWS ! Startup Developer J
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. AWS Cloud9
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. Analyzing ASK AN EXPERT Logs ! Tokenization Word2Vec
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. Analyzing ASK AN EXPERT Logs ! Tokenization from janome.tokenizer import Tokenizer t = Tokenizer("userdic.csv", udic_enc="utf8") f = io.open('./sodan.txt', 'r', encoding='utf-8’) tokens = t.tokenize(line) for token in tokens: partOfSpeech = token.part_of_speech.split(',')[0] if partOfSpeech == u'’: if token.surface == ‘https’: pass elif token.surface.isnumeric(): pass else: sodan_words.append(token.surface) https://github.com/mocobeta/janome
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. Analyzing ASK AN EXPERT Logs ! Word2Vec from gensim.models import word2vec sodan_sentences = word2vec.Text8Corpus('./sodan_words.txt') sodan_model = word2vec.Word2Vec(sodan_sentences, size=200, min_count=20, window=15) results = sodan_model.wv.most_similar(positive=[u'']) for result in results: print(result) https://github.com/RaRe-Technologies/gensim
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. ASK AN EXPERT Logs ! 43 9D 9 . 3 058675 2 3 1 . 3 - 2 3 43 058675 3 . 3 I 9D EAC :
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. ASK AN EXPERT Logs ! 2 A L . 201 65 - A 201 2 743 . 8 9:
© 2019, Amazon Web Services, Inc. or its Affiliates. All
rights reserved. AWS Loft Tokyo - ASK AN EXPERT Logs • EC2/RDS/S3 27)+&(70*T LO G;@ ! ⇒ <8=CU • "$Lambda'6,.FMIJT ?QK/ %524LOU ⇒ > J • AWS9SAWS(37-IJ:ND AB# (*´∀V*) " E HRAWS Loft Tokyo “ASK AN EXPERT”17*P J