Ontology

How to implement ontologies within your data pipeline

System Model

Figure 1: Non-Ontology Based (Traditional One) generic framework of Farm Animal Data Management

Non-ontology-based systems require rewriting the cleaning code each time new data arrives, making it time-consuming and tedious. These systems lack interoperability, flexibility, reusability, and accessibility, posing limitations in data integration, handling, and adaptability.

Figure 2: Ontology-Driven Knowledge-based framework of Farm Animal Data Management (ODKFADM)

In this use case, we utilize the ODKFADM framework to evaluate cattle-related data (DGZ), It involves collecting raw data from different data sources i.e. DGZ, and including information about farm identification, geo-location, infectious diseases, PCR, and bacterial culture results. This framework enables effective analysis of the data, leading to insights into cattle health

Step 1: Data Acquisition and RDF conversions

Figure 3: Data Acquisition and RDF Conversions

In Figure 3, we read heterogeneous raw data using pandas, and R data frames, and then convert them into RDF format.

Step 2: Species-Specific Ontology (LHO)

Figure 4: Graphical Representation of LHO

Step 3: RDF Data and Ontology Integration (Mapping)

In this step we map the RDF data with LHO which enhances the reasoning and query capabilites

Figure 5: RDF and Ontology Integration (Mapping)

Step 4: Knowledge graph (Ontology Update)

Resulted into a knowledge graph

Step 5: Reasoning and Query

Figure 6 and 7 shows Query and query results that Filtering positive PathogenResults and MycoplasmaResults associated with CattleSample

Figure 6: Filtering positive PathogenResults and MycoplasmaResults associated with CattleSample

Figure 7: Query result

Step 6: Visualization and Analysis

We choose Tableau for the visualization method. It provides meaningful insights to explore the knowledge graph. For this, we need a working ODBC connection to a Virtuoso Instance and ODBC or JDBC Compliant version of Tableau or Tableau Server. For ODBC connection to virtuoso, the link is: Visualizing SPARQL Results in Tableau | by Daniel Heward-Mills | OpenLink Virtuoso Weblog | Medium