Thursday 6 October 2022

A Graph Database and a Dadjoke walk into a bar...

I just publised a blogpost series with 6 different articles about me having fun with Dadjokes, in an unusual sort of way. Here are the links to the articles:

All of the queries etc are put together in this markdown document. I plan to make a Neo4j Guide out of this as well in the next few days so that it would become easier to use. 

Hope you will have as much fun with it as I did. 


DadjokeGraph Part 6/6: Closing: some cool Dadjoke Queries

A Graph Database and a Dadjoke walk into a bar...

Now that we have a disambiguated graph of dadjokes, let's have some fun and explore it.

How many times does a joke get tweeted?

WITH dj.Text AS Joke, count(r) AS NrOfTimesTweeted
RETURN Joke, NrOfTimesTweeted
ORDER BY NrOfTimesTweeted DESC

How many times does a joke get favorited?

RETURN dj.Text AS Joke, dj.SumOfFavorites AS NrOfTimesFavorited, dj.SumOfRetweets AS NrOfTimesRetweeted
ORDER BY NrOfTimesFavorited DESC

DadjokeGraph Part 2/6: Importing the Dadjokes into the Dadjoke Graph

A Graph Database and a Dadjoke walk into a bar...

This means that we want to convert the spreadsheet that we created before, or the .csv version of it, into a Neo4j Database.

Here's how we go about this. First things first: let's set up the indexes that we will need later on in this process:

CREATE INDEX tweet_index FOR (t:Tweet) ON t.Text;
CREATE INDEX dadjoke_index for (d:Dadjoke) ON d.Text;

Assuming the .csv file mentioned above is in the import directory of the Neo4j server, we can use load csv to create the initial dataset:

LOAD CSV WITH HEADERS FROM "file:/vicinitas_alldadjokes_user_tweets.csv" AS csv
CREATE (t:Tweet)
SET t = csv;

Import the Vicinitas .csv file

Or: if you want to create the graph straight from the Google Sheet:

CREATE (t:Tweet)
SET t = csv;

DadjokeGraph Part 3/6: Taking on the real disambiguation of the jokes

A Graph Database and a Dadjoke walk into a bar...

We noticed that many of the Tweet nodes referred to the same jokes - and resolved that already above. But this query makes us understand that we actually still have some work to do:

MATCH path = (dj:Dadjoke)-[*..2]-(conn)
WHERE dj.Text CONTAINS "pyjamazon"
    RETURN path;

The Amazon Dadjokes

We will come back to that example below.

We now notice that there are quite a few Dadjoke nodes that are a bit different, but very similar. We would like to disambiguate these too. We will use a couple of different strategies for this, but start with a strategy that is based on String Metrics.

DadjokeGraph Part 4/6: Adding NLP and Entity Extraction to prepare for further disambiguation

A Graph Database and a Dadjoke walk into a bar...

As we can see in the pyjamazon example from before, the disambiguation of our Dadjokes has come a long way - but is not yet complete. Hence we we call the graph to the rescue here, and take it a final step further that will provide a wonderfully powerful example of how and why graphs are so good at analysing the structural characteristics of data, and make interesting and amazing recommendations on the back of that.

Here's what we are going to do:

  1. we are going to use Natural Language Processing to extract the entities that are mentioned in our Dadjokes. Do do that, we are going to use the amazing Google Cloud NLP Service, and call it from APOC. This will yield a connected structure that will tell us exactly which entities are mentioned in every joke.
  2. then we are going to use that graph of dadjokes connected to entities to figure out if the structure of the links can help us with further disambiguation of the jokes.

So let's start with the start.

DadjokeGraph Part 5/6: Disambiguation using Graph Data Science on the NLP-based Entities

A Graph Database and a Dadjoke walk into a bar...

The next, and final, step our dadjoke journey here, is going to be taking the disambiguation to the next level by applying Graph Data Science metrics to the new, enriched (using NLP), structure that we have in our graph database. The basic idea here is that, while the TEXT similarity of these jokes may be quite far apart, their structural similarity may still be quite high based on the connectivity between the joke and its (NLP based) entities.

Calculating the Jaccard similarity metric using GDS

To explore this, we will be using the Jaccard similarity coefficient, which is part of the Neo4j Graph Data Science library that we have installed on our server. More about this coefficient can be found on Wikipedia. The index is defined as the size of the intersection divided by the size of the union of the sample sets, which is very well illustrated on that Wikipedia page. I have used Neuler (the no-code graph data science playground that you can easily add to your Neo4j Desktop installation) to generate the code below - but you can easily run this in the Neo4j Browser as well.

DadjokeGraph Part 1/6: Building a Dadjoke database - from nothing

A Graph Database and a Dadjoke walk into a bar...

I am a dad. Happily married and 3 wonderful kids of 13, 17 and 19 years old. So all teenagers - with all the positive and not so positive experiences that can come with that. And in our family, I have been accused of using dadjokes - preferably at very awkward or inconvenient times. I actually like dadjokes. A lot.

So I follow a few accounts on social media that post these jokes. Like for example baddadjokes, dadjokeman, dadsaysjokes, dadsjokes, groanbot, punsandoneliner, randomjokesio, and thepunnyworld and there are many others. These are all in this list, should you be interest. It's a very funny list.

Dadjokers List on Twitter