开始一个 GitHub 代码库的十个创意，助您打入数据科学领域。.docx

10IDEASFORSTARTINGAGITHUBPORTFOLIOTOBREAKINTODATASCIENCE

MIGUELFIERRO

Copyright(c)MiguelFierro.Allrightsreserved.

Theinformationandfilescontainedhereareconfidential.Neitherthisdocumentnortheinformationcontainedhereinmaybe,inwholeorinpart,published,reproduced,copied,disseminated,ordistributedinanywaywithouttheexpress,prior,writtenpermissionofMiguelFierro.

v1.1

Welcome!

Welcome!I’mMiguelFierro.

Youmightalreadyknow,butcompaniesthesedaysaremoreinterestedinpracticalknowledgethanintheoreticalknowledge.

That’swhyCVsarelessandlessrelevant,your

onlinepresenceisthebestCV.

That’swhycreatingaDataScienceportfoliois

thebestwaytogetajobinDataScience.

Next,you’llfind10ideastobuildyourportfolio.

Vamos!

HATESPEECHDETECTION

Description:

SentimentanalysisofTweetstoidentifywhetherthetextishatespeech,offensivelanguage,orneutral.

Dataset:

KaggleHateSpeechandOffensiveLanguageDataset

Library:

HuggingfaceTransformers

Algorithm:

DeBERTa

FORECASTINGOFCOVID

Description:

Usetime-seriesforecastingtopredictthe

spreadofCOVIDindifferentcountries.

Dataset:

KaggleCOVID-19datasetorGoogleCOVID-19OpenData

Library:

MetaAIProphet

Algorithm:

Prophet

MOVIERECOMMENDATION

Description:

Recommendamoviebasedonhistoricaluser

behavior.

Dataset:

Movielens

Library:

MicrosoftRecommenders

Algorithm:

SAR

CARTOONCLASSIFICATION

Description:

IdentificationoftheSimpsoncharactersin

images

Dataset:

KaggleTheSimpsonsCharactersData

Library:

Torchvision(PyTorch)

Algorithm:

ResNet18

FOOTBALLPLAYERTRACKING

Description:

Playertrackinginfootballmatchesfor

obtainingmatchmetrics.

Dataset:

SportsVideosintheWildfromMichiganUniversity

Library:

UltralyticsYOLOv5

Algorithm:

YOLOv5

NEWSRECOMMENDATION

Description:

RecommendationofnewsarticlesmixingNLP

andRecommendationSystems.

Dataset:

MicrosoftMINDdataset

Library:

MicrosoftRecommender