book

Network Science with Python

by David Knickerbocker

February 2023

Intermediate to advanced

414 pages

9h 24m

English

Packt Publishing

Read now

Unlock full access

AcknowledgementsContributorsAbout the authorAbout the reviewers
Who this book is forWhat this book coversDownload the example code filesConventions usedGet in touchShare Your ThoughtsDownload a free PDF copy of this book
Technical requirementsWhat is NLP?Why NLP in a network analysis book?A very brief history of NLPHow has NLP helped me?Simple text analysisCommunity sentiment analysisAnswer previously unanswerable questionsSafety and securityCommon uses for NLPTrue/False – Presence/AbsenceRegular expressions (regex)Word countsSentiment analysisInformation extractionCommunity detectionClusteringAdvanced uses of NLPChatbots and conversational agentsLanguage modelingText summarizationTopic discovery and modelingText-to-speech and speech-to-text conversionMTPersonal assistantsHow can a beginner get started with NLP?Start with a simple ideaAccounts that post most frequentlyAccounts mentioned most frequentlyTop 10 data science hashtagsAdditional questions or action items from simple analysisSummary
The confusion behind networksWhat is this network stuff?Graph theorySocial network analysisNetwork scienceResources for learning about network analysisNotebook interfacesIDEsNetwork datasetsKaggle datasetsNetworkX and scikit-network graph generatorsCreating your own datasetsNetworkX and articlesCommon network use casesMapping production dataflowMapping community interactionsMapping literary social networksMapping historical social networksMapping languageMapping dark networksMarket researchFinding specific contentCreating ML training dataAdvanced network use casesGraph MLRecommendation systemsGetting started with networksExample – K-pop implementationSummaryFurther reading
Technical requirementsUsing notebooksData analysis and processingpandasNumPyData visualizationMatplotlibSeabornPlotlyNLPNatural Language ToolkitSetupStarter functionalityDocumentationspaCyNetwork analysis and visualizationNetworkXscikit-networkMLscikit-learnKarate ClubspaCy (revisited)Summary
Technical requirementsWhy are we learning about NLP in a network book?Asking questions to tell a storyIntroducing web scrapingIntroducing BeautifulSoupLoading and scraping data with BeautifulSoupChoosing between libraries, APIs, and source dataUsing NLTK for PoS taggingUsing spaCy for PoS tagging and NERSpaCy PoS taggingSpaCy NERConverting entity lists into network dataConverting network data into networksDoing a network visualization spot checkAdditional NLP and network considerationsData cleanupComparing PoS tagging and NERScraping considerationsSummary
Technical requirementsWhy cover Requests and BeautifulSoup?Introducing Newspaper3kWhat is Newspaper3k?What are Newspaper3k’s uses?Getting started with Newspaper3kScraping all news URLs from a websiteScraping a news story from a websiteScraping nicely and blending inConverting text into network dataEnd-to-end Network3k scraping and network visualizationIntroducing the Twitter Python LibraryWhat is the Twitter Python Library?What are the Twitter Library’s uses?What data can be harvested from Twitter?Getting Twitter API accessAuthenticating with TwitterScraping user tweetsScraping user followingScraping user followersScraping using search termsConverting Twitter tweets into network dataEnd-to-end Twitter scrapingSummary
Technical requirementsCreating a graph from an edge listTypes of graphsSummarizing graphsListing nodesRemoving nodesQuick visual inspectionAdding nodesAdding edgesRenaming nodesRemoving edgesPersisting the networkSimulating an attackSummary

Technical requirementsCreating baseline WNA questionsRevised SNA questionsSocial network analysis revisitedWNA in actionLoading data and creating networksNetwork size and complexityNetwork visualization and thoughtsImportant nodesDegreesDegree centralityBetweenness centralityCloseness centralityPageRankEdge centralitiesComparing centralitiesVisualizing subgraphsInvestigating islands and continents – connected componentsCommunitiesBridgesUnderstanding layers with k_core and k_coronak_corek_coronaChallenge yourself!Summary
Technical requirementsEgocentric network analysisUses for egocentric network analysisExplaining the analysis methodologyInvestigating ego nodes and connectionsEgo 1 – ValjeanEgo 2 – MariusEgo 3 – GavrocheEgo 4 – JolyInsights between egosIdentifying other research opportunitiesSummary
Technical requirementsIntroducing community detectionGetting started with community detectionExploring connected componentsUsing the Louvain methodHow does it work?The Louvain method in action!Using label propagationHow does it work?Label propagation in action!Using the Girvan-Newman algorithmHow does it work?Girvan-Newman algorithm in action!Other approaches to community detectionSummary
Technical requirementsIntroducing MLBeginning with MLData preparation and feature engineeringDegreesClusteringTrianglesBetweenness centralityCloseness centralityPageRankAdjacency matrixMerging DataFramesAdding labelsSelecting a modelPreparing the dataTraining and validating the modelModel insightsOther use casesSummary
Technical requirementsWhat is unsupervised ML?Introducing Karate ClubNetwork science optionsUses of unsupervised ML on network dataCommunity detectionGraph embeddingsConstructing a graphCommunity detection in actionSCDGraph embeddings in actionFEATHERNodeSketchRandNEOther modelsUsing embeddings in supervised MLPros and consLoss of explainability and insightsAn easier workflow for classification and clusteringSummary
Other Books You May EnjoyPackt is searching for authors like youShare Your ThoughtsDownload a free PDF copy of this book

Content preview from Network Science with Python

6 Graph Construction and Cleaning

We have covered quite a lot of ground up to this point. In the previous chapters, we introduced NLP, network science, and social network analysis, and we learned how to convert raw text into network data. We even visualized a few of these networks. I hope that seeing text converted into a visualized network had the same impact on you as it did on me. The first time I attempted this, I used the book of Genesis, from the Bible, and being able to convert text from thousands of years ago into an actual interactive network took my breath away.

In the previous two chapters, we learned a few different ways to collect text data from websites and social networks on the internet and to use that text data to create networks. ...