HowTo -- Generating Instant DBpedia Lookups and Interactions using ChatGPT

[kidehen]

Situation Analysis

When it comes to structured data querying, lookups, and access, CHAT-GPT can provide instant access if the data in question is deployed using Linked Data principles i.e., entity and entity relationship type naming using hyperlinks (clickable identifiers) that manifest a Semantic Web variant of a Knowledge Graph.

The transcript that follows comprises prompts that demonstrate the point outlined above; and for additional assistance, I’ve also included a link to a YouTube screencast that showcases actions performed during the process.

Transcript

Q: List 10 clickable identifiers from DBpedia that denote classes

A:

DBpedia is a project that extracts structured content from the information available in Wikipedia. It allows users to query relationships and properties associated with Wikipedia resources, including links to other related datasets. The following is a list of 10 clickable identifiers from DBpedia that denote classes:

1. dbo:Person - Represents a human being.

2. dbo:Place - Represents a geographical location or an entity with a physical extent.

3. dbo:Organisation - Represents an organized group of people, such as companies, institutions, or agencies.

4. dbo:Event - Represents an occurrence, happening, or organized gathering.

5. dbo:Work - Represents creative works, such as books, films, or musical compositions.

6. dbo:Species - Represents a group of living organisms with common characteristics.

7. dbo:Activity - Represents actions, tasks, or hobbies performed by humans or other entities.

8. dbo:Software - Represents computer programs and applications.

9. dbo:Device - Represents physical objects or tools used by humans or other entities.

10. dbo:Award - Represents an official recognition or honor given to a person, organization, or work.

Q: Given the following URL pattern: https://dbpedia.org/describe/?uri=, generate clickable links using each DBpedia class identifier.

A:

Here are clickable links for each DBpedia class identifier, using the provided URL pattern:

1. dbo:Person - Represents a human being.

2. dbo:Place - Represents a geographical location or an entity with a physical extent.

3. dbo:Organisation - Represents an organized group of people, such as companies, institutions, or agencies.

4. dbo:Event - Represents an occurrence, happening, or organized gathering.

5. dbo:Work - Represents creative works, such as books, films, or musical compositions.

6. dbo:Species - Represents a group of living organisms with common characteristics.

7. dbo:Activity - Represents actions, tasks, or hobbies performed by humans or other entities.

8. dbo:Software - Represents computer programs and applications.

9. dbo:Device - Represents physical objects or tools used by humans or other entities.

10. dbo:Award - Represents an official recognition or honor given to a person, organization, or work.

Q: Generate a Table comprising columns DBpedia ID and Faceted Browser Page using the clickable links denoting classes from each of the prior requests.

Q: Redo, since you didn’t generate clickable links for values in the DBpedia column. Also output to a Markdown table using a code-block

A:

DBpedia ID	Faceted Browser Page
dbo:Person	Person
dbo:Place	Place
dbo:Organisation	Organisation
dbo:Event	Event
dbo:Work	Work
dbo:Species	Species
dbo:Activity	Activity
dbo:Software	Software
dbo:Device	Device
dbo:Award	Award

Related

• Screencast Demonstrating Instant DBpedia Lookups using CHAT-GPT

• CHAT-GPT Transcript

[mdebellis]

Just FYI: the results that ChatGPT provides don’t always work. IMO that is one of the biggest drawbacks with ChatGPT, even when it doesn’t know it still states results with confidence (maybe it is self aware, certainly behaving too much like a human in that sense). I’ve tried several times and at first was psyched by the query ChatGPT returned but then when I tried it it didn’t work. I’m not saying not to use it at all, just saying don’t assume that because it churns out a query that the query will work.

Also, just some feedback from a new user… actually an ex new user, I find the documentation and help on DBpedia to be far below other tools. I posted 2 questions on what I thought was supposed to be the user forum and never got a reply to either and I searched in vain for some basic documentation on the structure of the ontology, etc. and came up empty. On the advice of someone I tried Wikidata instead and it is far superior. In one hour this morning I learned far more than I was able to in many hours of research on DBpedia.

[kidehen]

@mdebellis

Just FYI: the results that ChatGPT provides don’t always work. IMO that is one of the biggest drawbacks with ChatGPT, even when it doesn’t know it still states results with confidence (maybe it is self aware, certainly behaving too much like a human in that sense).

You can achieve more predictability using GPT4. My examples are GPT4 based unless I indicate otherwise.

[kidehen]

@mdebellis

I find the documentation and help on DBpedia to be far below other tools. I posted 2 questions on what I thought was supposed to be the user forum and never got a reply to either and I searched in vain for some basic documentation on the structure of the ontology,

The term “DBpedia” remains a subtle problem i.e., it is an ambiguous terms that can refer to so many different things.

For instance, do you have a link to the documentation in question?

The DBpedia Ontology is best viewed using a dump in your preferred format or by just browsing it using its Faceted Search & Browsing interface.