The last decade has seen an enormous growth of social media platforms, such as Facebook, Twitter, and Youtube. Since 2009, another platform with a different format and objective has grown - Pinterest. Unlike conventional social media, which is used as a communication tool, Pinterest is described as a "catalog of ideas". It allows users to create pinboards, and to organize and share content over the web, based on interests and ideas.

We've explored the use of the Pinterest API and also advanced scraping techniques, using Selenium and BeautifulSoup, to gather data for learning purposes. However, these have time constraints to be scalable. Therefore, we extracted the data from our own pinboard using the endpoints (user, board, and pins) and the search results on the topic of fashion. For data analysis, we used bigram analysis to extract a list of topics on our own pins and then visualized it in a graph structure using NetworkX (via fruschterman_reingold_layout). We then studied centrality metrics and four measures: degree, closeness, betweenness, and eigenvector. We applied these measures to our pinboard and found some interesting results. Then we extracted the topics on the data gathering from the search results on Pinterest on the subject of fashion. We used bigram analysis to discover some interesting results from trends, events, style, and clothes.

The next section was dedicated to finding user relationships on shared interests and topics, through graph relationships and community detection. We achieved this by first building a graph of the user network based on the similarity of pinned topics between all the users. Then we used the centrality measures to find out the most connected, central, and influential users. Finally, we provided an introduction to community detection using Python's community module, and visualized the community of users that shared similar sub-topics. We touched upon various important concepts including advanced data scraping, graph theory and visualization, centrality, and community detection.

These exciting concepts and skills lead us to the next and last chapter, Social Media Analytics at Scale, where we'll learn about several topics around distributed and parallel computing, Spark and Amazon Web Services. These will take your social media analytics skills to the next level!