资源描述：
《setting the pace what is bad data》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。
1、CHAPTER1SettingthePace:WhatIsBadData?Weallsaywelikedata,butwedon’t.Welikegettinginsightoutofdata.That’snotquitethesameaslikingthedataitself.Infact,IdaresaythatIdon’tquitecarefordata.ItsoundslikeI’mnotalone.It’stoughtonaildownaprecisedefinitionof“BadData.”Somepeopleconsideritapurelyhands-
2、on,technicalphenomenon:missingvalues,malformedrecords,andcran‐kyfileformats.Sure,that’spartofthepicture,butBadDataissomuchmore.Itincludesdatathateatsupyourtime,causesyoutostaylateattheoffice,drivesyoutotearoutyourhairinfrustration.It’sdatathatyoucan’taccess,datathatyouhadandthenlost,data
3、that’snotthesametodayasitwasyesterday…Inshort,BadDataisdatathatgetsintheway.Therearesomanywaystogetthere,fromcrankystorage,topoorrepresentation,tomisguidedpolicy.Ifyoustickwiththisdatasciencebitlongenough,you’llcertainlyencounteryourfairshare.Tothatend,wedecidedtocompileBadDataHandbook,a
4、roguesgalleryofdatatrou‐blemakers.Wefound19peoplefromallreachesofthedataarenatotalkabouthowdataissueshavebittenthem,andhowthey’vehealed.Inparticular:GuidanceforGrubby,Hands-onWorkYoucan’tassumethatanewdatasetiscleanandreadyforanalysis.KevinFink’sIsItJustMe,orDoesThisDataSmellFunny?(Chapt
5、er2)offersseveraltechniquestotakethedataforatestdrive.There’splentyofdatatrappedinspreadsheets,aformatasprolificasitisincon‐venientforanalysisefforts.InDataIntendedforHumanConsumption,NotMachineConsumption(Chapter3),PaulMurrellshowsoffmovestohelpyouextractthatdataintosomethingmoreusable.
6、1Ifyou’reworkingwithtextdata,soonerorlateracharacterencodingbugwillbiteyou.BadDataLurkinginPlainText(Chapter4),byJoshLevy,explainswhatsortofproblemsawaitandhowtohandlethem.Towrapup,AdamLaiacano’s(Re)OrganizingtheWeb’sData(Chapter5)walksyouthrougheverythingthatcangowronginaweb-scrapingeff
7、ort.DataThatDoestheUnexpectedSure,peoplelieinonlinereviews.JacobPerkinsfoundoutthatpeoplelieinsomeverystrangeways.TakealookatDetectingLiarsandtheConfusedinContradictoryOnlineReviews(Chapter6)tolearnhowJacob’snatural-languageprogramming(NLP)workuncoveredthisnewbreedoflie.O
