The author used Neo4j, a graph database, to model and analyze data from the Titanic's passenger manifest. They started by creating nodes for each passenger, relationship connections between them, and properties of relationships such as ticket type and embarkation point. However, they soon realized that categorizing the data differently was necessary to simplify queries and better understand the social implications of the data. The author restructured their graph to make more data into properties of nodes, allowing for easier querying and analysis. They extracted specific data from the CSV, such as cabin and boat side information, and used it to answer questions about survivorship based on factors like gender, age, and side of the ship assigned a cabin or lifeboat. The analysis revealed that females had a higher survival rate than males, children had a very high survival rate regardless of their gender, and there was little difference in survivorship related to which side of the ship each passenger was assigned a cabin. Additionally, the data suggested that adult males who tried to escape from the port side did so in boats with limited capacity. The author emphasizes the importance of continually investigating and augmenting data to gain deeper insights into complex social issues.