I just did a search on “Neo4j” on the forum and came across your question. Sorry for the late reply!
My approach is to query the relevant tables in Airtable through its API. The API only permits 100 rows at a time so you’ll need to detect an offset and keep pulling down the data. Once you have pulled down all the relevant data, you can then (in a first phase) creates nodes in Neo4j and (in a second phase) connect them.
If all the data is in a single base, you can just Airtable’s own record ids as Neo4j’s ids, which makes it easy to connect nodes with the relationships between the data in your Airtable tables.
Pretty much any language can be used to export the data from Airtable. If you’re not comfortable with coding, you have the option of copying and pasting columns into a text file and then importing them into Neo4j as a CSV file. In fact, when you have millions of rows of data, CSV is the fastest import option.
I regularly move my own Airtable data into Neo4j and its a great way to query and visualize patterns in your data. Where Airtable and Neo4j both shine is in making your data “touchable”, which leads to a much higher quality of data because you’re living with it.
Do reply if you found a solution.