parkererickson / graphAIWorldDataframeToGraph

Repository containing the code for Parker Erickson's Graph+AI World Presentation

Enabling Data Science with pyTigerGraph

Code for the Graph+AI World 2020 Conference Presentation

This repository contains the code for the Graph+AI World session From Dataframes to Graph: Data Science with pyTigerGraph by Parker Erickson.

Software Compatibility

Unfortunately, this code will only work on UNIX systems, and they must have Java installed for the Python GSQL interface. This is something we are trying to remedy for future versions of pyTigerGraph.

Download Data

Download the data here. The directory structure should be:

graphAIWorldDataframeToGraph
                            |-data
                            |     |- artistData.tsv
                            |     |_ etc.
                            |     
                            |-createSchemaAndLoad.ipynb
                            |-genreClassification.ipynb
                            |_ ...

Create Free tgcloud Account

The notebooks assume that you are using a tgcloud instance of TigerGraph. You can create a free instance, and choose "Blank" for the starter kit.

Python Packages

This repository uses Pandas, pyTigerGraph, Pytorch, and DGL (might be missing a few there). Make sure you have these installed before trying to use the notebooks.

Graph Machine Learning Resources

Building out this site with many different papers and videos to get you started with graph machine learning.