The author has been working on linking resident records from the 1901 and 1911 Irish censuses together to identify the same residents. The data model used for the imported data is linear, with houses and residents as nodes. To improve record linkage scoring, the author has developed a method that uses local data space around objects to match records. This involves creating relationships between residents based on their properties and using graph-based comparisons to evaluate family relationships. The author has implemented this method using Neo4j, a powerful tool for analyzing data relationships naturally and quickly. By leveraging Neo4j's ability to create relationships on the fly, the author is able to find the best solution quickly and efficiently.