Case Study: Kode Solutions achieves semantic tweet analytics and graph-powered social insights with ArangoDB

Meet Kode Solutions Semantic Tweet Analytics with ArangoDB & Rlang

Kode Solutions is a machine learning agency from Pisa and Cagliari that needed a scalable way to do semantic tweet analytics—monitoring political discussions, extracting users/hashtags from Twitter JSON and turning them into a knowledge graph for analysis. To build their Tweetmood engine they integrated ArangoDB and a custom R connector (aRangodb) so analysts could ingest tweets and users, model relationships, and run analytics without heavy engineering overhead.

The solution used ArangoDB as a graph database with three collections (users, tweets, hashtags) and three edge types (wrote, mentioned_in, contained_in), orchestrated by Tweetmood microservices and the aRangodb package. Using ArangoDB enabled association-rule and co‑occurrence analyses that produced high-confidence rules (examples include rules with confidence up to 1.0, lifts >30 and rule counts in the hundreds—e.g., 977 and 654 occurrences), and allowed Kode Solutions to persist itemsets and rules as graph data; ArangoDB thus directly improved their ability to discover actionable semantic insights from Twitter.

Kode Solutions

Gabriele Galatolo

Data Scientist & Software Developer