Developing a knowledge graph application

Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

Project 2 -- Advice

Steps in developing a knowledge graph application

1. Requirements analysis: Analyze:

· relevant content (text, current applications/interfaces) – to identify entities, concepts, relations, functions (actions)

· users and purpose of application -- how the application helps the user

· intelligent application

2. Look for existing data in spreadsheet format or database format.

3. Construct taxonomies/class hierarchies, relations between concepts, relation hierarchy.

4. Represent the concepts and instances (and their properties) as spreadsheet tables. These will be represented in the knowledge graph as nodes (with properties)
Also represent relations (links) between instances as spreadsheet tables. These will be represented in the knowledge graph as edges (links and arrows).

5. Design the graph database (we’re using Neo4j) -- by constructing the Cypher commands for importing the CSV files into a Neo4j graph database.

6. Design the queries for retrieving data, and inferring new properties and relations to satisfy the purposes of the application and the intelligent application (identified in the requirements analysis). The query results are used to display information to help the users, and also to implement the intelligent applications.

7. Design the output (results) of the Cypher queries -- to be user friendly. Design the network visualization of the results (using Neo4j browser), including screenshots of your network visualizations.

General Comments on Project 2 Reports

General

The report should have a teaching function:

· Novelty – communicates new information to the reader

· Audience – you have to decide on the target reader, e.g. your boss, the KM manager who will take over the project

· Innovation – shows that you have developed or found out something new, that others can learn from you.

· Coherent – integrates information to synthesize new knowledge or new understanding for the reader

A grade:  80 marks and above

A-  :  75-79

B+ :  70-74

B   :  65-69

Don’t

Don’t describe the steps in using the software. The reader can easily go read the software manual.

Unless: you found out something that will help the reader to avoid your mistakes or problems.

Don’t describe the detailed steps in ontology or knowledge graph construction. (Just briefly outline the steps.)

Unless: you have an innovative step. Or you focused on one of the steps – and it affects your ontology construction.

Don’t describe your learning journey, and the many problems and challenges you faced (e.g., software license expiring).

Unless: it helps the reader – in alerting of unexpected problems, and how to avoid or address those problems.

Don’t list your taxonomy or ontology in the body of your report (the 6 pages). And don’t describe your ontology in detail in the text. It is boring and reader can’t learn much.

Do: point out the overall structure of the ontology, any design principles you used, interesting or new characteristics of the ontology.

Do

Do think of more intelligent applications, and suggest detailed steps (e.g., Cypher queries) in implementing the intelligent applications.