Introduction to graph databases

This talk presents differences between SQL and NoSQL databases and explains the efficiency of graph databases. The main features and Neo4j use cases are defined as well as a basis for building a prototype.
This presentation is used for companies consulting on Neo4j.

1. Introduction to graph databases Roksolana Diachuk, WWCode Kyiv Lead, Data Engineer at Ciklum
2. Relational vs. non-relational • Relational DBs store data in tables and rows. NoSQL DBs store data in various formats - documents, graphs • Relational DBs use unified query language - SQL. NoSQL DBs may have various query languages. • Relational DBs provide ACID guarantees that is why they are more reliable. NoSQL DBs usually do not have strict guarantees (they may provide BASE guarantees) • It is easier to scale NoSQL DBs and change the schema. While in relational DBs the schema is fixed. @dead_flowes22
3. Definition Graph database is NoSQL database which is based on the graph theory. All the data is expressed with the help of nodes (entities) and edges (relationships between entities). @dead_flowes22
4. Why would you care about graph databases • Flexibility (for volume increase) • High performance • Agility (for schema) @dead_flowes22
6. Neo4j query @dead_flowes22
7. Neo4j Neo4j is one of the most popular graph databases. The query language is Cypher which is similar to SQL. Neo4j management: 1. Cypher shell 2. Admin 3. REST API @dead_flowes22
8. @dead_flowes22
9. Neo4j queries visualization @dead_flowes22
10. @dead_flowes22
11. Real-time recommendation systems @dead_flowes22
12. Fraud detection @dead_flowes22
13. Master data management @dead_flowes22
14. Graphs aided search @dead_flowes22
15. Compliance and privacy solutions @dead_flowes22
16. Prototype creation stages • Dataset research • Database design (edges and relationships) • Dataset import • Queries on dataset @dead_flowes22
17. Conclusion Building a prototype with graph databases will help you to discover how graph databases can be integrated into your system and whether your data can be queried and managed more efficiently. @dead_flowes22
18. Thank you for attention!