狠狠撸

狠狠撸Share a Scribd company logo
Being a Unicorn in Data Science
through Problem Solving
Yao?JenKuo
November1st2017
關於我
var?yaoJenKuo?=?{?
????"name":?"郭耀仁",?
????"organization":?"Kyosei.ai",?
????"loves":?[?
????????"Data?Science",?
????????"Marathon",?
????????"Ping?pong"?
????]?
}??
[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving
台大資工系統訓練班(2016?現在)
國立台北商業大學(2017秋)
玉山商業銀行(2017夏)
Could that data BE any tidier?
骋辞辞驳濒别图片
骋辞辞驳濒别图片
DataScientist:TheSexiestJobofthe21st
Century.
HarvardBusinessReviewOCTOBER2012ISSUE
骋辞辞驳濒别图片
baa.org
Unicorn
Ibelievetheworldcouldbeabetterplace
ifmoredesignerslearnedhowtocode
andmoredeveloperslearnedhowto
design.Theyaretheoneswhowill
experimentandplay.Theyaretheones
whowilltakerisks.Theyaretheoneswho
willbuild&launchtheirideas.
MackenzieChild–Medium
DataHacker
QuotedfromAjaiSharma@LinkedIn
Getdirtywithdata,technologyandmath,
nothingshouldbetoopracticalenough,
tootheoreticalenough,ortoo
mathematicalenoughforyoutohandle.
10xDataScientist(誤)
骋辞辞驳濒别图片
10xDataScientist
寫程式
數學、統計與機率
機?学习
永遠抱持懷疑
擅長溝通
Quora
骋辞辞驳濒别图片
潛在的不效率分工與坑
Dataanalyst
Dataengineer
Datascientist
[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving
管理者如何解决这个困境
骋辞辞驳濒别图片
骋辞辞驳濒别图片
骋辞辞驳濒别图片
Photo?via?via?VisualHunt.com
理想的閱讀時間是工作日的午餐時間,我認
為最好的学习就是不佔用讀者工作以外的休
憩時間。
輕鬆学习R語言
成员如何解决这个困境
骋辞辞驳濒别图片
骋辞辞驳濒别图片
骋辞辞驳濒别图片
学习
骋辞辞驳濒别图片
常見資料科學問題
资料来源
资料处理
资料视觉化
機?学习
高校能计算
资料来源
表格式文字檔
#?R?
?read.table?
#?Python?
import?pandas?as?pd?
?pd.read_table?
Excel試算表
#?R?
?readxl::read_excel?
#?Python?
import?pandas?as?pd?
?pd.read_excel?
JSON
JSONtypes R Python
ArrayofJSON data.frame DataFrame
JSONobject list dict
#?R?
?jsonlite::fromJSON?
#?Python?
import?pandas?as?pd?
import?json?
?pd.read_json?
?json.loads?
資料庫
#?R?
library(DBI)?
?
con?<??dbConnect(RMySQL::MySQL(),??
?dbname?=?"friends",?
?host?=?"127.0.0.1",?
?port?=?3306,?
?user?=?"chandlerbing",?
?password?=?"muriel")?
dbDisconnect(con)?
資料庫(2)
#?Python?
import?mysql.connector?
?
con?=?mysql.connector.connect(user='chandlerbing',?password='muriel',?
??????????????host='127.0.0.1',?
??????????????database='friends')?
con.close()?
網頁資料入門
W3Schools
SelectorGadget
XPathHelper
rvest
beautifulsoup
網頁資料進階
Scrapy
SeleniumwithPython
RSelenium:Basics
资料处理
[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving
用適當的資料結構儲存資料
維度
變數類型
對稱性
资料视觉化
Photo?credit:? ?via? ?/?Horia?Varlan VisualHunt CC?BY
靜態圖表
PythonGraphGallery
RGraphGallery
具有互動元件的動態圖表
BokehGallery
ShinyGaller
建立与维护作品集
[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving
[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving
輕量CMS
Jekyll
Hugo
Hexo
[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving
機?学习
複習線性代數
http://www.numpy.org/
複習統計與機率
http://www.statsmodels.org/stable/index.html
[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving
高校能计算
SeamlessRandC++Integration
http://www.rcpp.org/
C?ExtensionsforPython
http://cython.org/
Spark
Scala
PySpark
sparklyr
Julia
AsfastasC,aseasyasPython,asgood
atstatisticsasR,asgoodasstring
processingasPerl,asscalableasHadoop
orR.
ViralShah,co?creatorofJulia
Docendo discimus
教學相長
Meetups
Udemy
[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving
[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving
路途很長
毋免驚,咱寬寬仔行。
滅火?:?上的人
? Contact:tony@kyosei.ai
Thanks for Watching
http://kyosei.ai ? DownloadPDF

More Related Content

[2017 臺灣資料科學年會]Being a Unicorn in Data Science through Problem Solving