Photo from Unsplash by Janis Fasel

Linear Regression with a Graph Database

How to Run a Linear Regression using Pokémon Data from a Graph in TigerGraph

(Note: This is a bonus blog based on a series. To obtain the data used, refer to the past blogs.)

Hello! Today, we’re going to learn how to run a linear regression using TigerGraph. In this example, we’ll be running a linear regression with the Pokémon’s height and weight. Let’s get to it!

Part I: Setup

To start, we’re going to need to import the necessary libraries and create a connection to our graph.

The libraries we’ll be using are pyTigerGraph, Plotly Express, and Pandas. You can import them using the following:

import pyTigerGraph as tg
import plotly.express as px
import pandas as pd

Note that if you are using a Colab notebook, you’ll need to install pyTigerGraph first:

!pip install pyTigerGraph

Now that you have your libraries installed, start your solution on TigerGraph Cloud and create a connection in Python.

conn = tg.TigerGraphConnection(host="https://SUBDOMAIN.i.tgcloud.io", password="tigergraph", gsqlVersion="3.0.5", useCert=True)
conn.graphname = "PokemonGraph"
conn.apiToken = conn.getToken(conn.createSecret())

Step II: Write a Query

Next, write a query to grab all the Pokémon. Here, we’ll just create a seed with all the Pokemon vertices and use that. We’ll assign the results of the query to a variable called result.

result = conn.runInterpretedQuery('''
INTERPRET QUERY () FOR GRAPH PokemonGraph {
catchThemAll = {Pokemon.*}; # Pokemon.* means that it'll grab all the Pokemon
PRINT catchThemAll; # print displays the results
}
''')

Step III: Create a Dataframe

To easily plot the data into Plotly, we’re going to need to format our results into a dataframe. Here’s how we’d do that:

result = result[0]["catchThemAll"] # Grabbing just the catchThemAll result
# The lists of what we need
height = []
weight = []
for i in result: # For each pokémon
    height.append(i["attributes"]["height"]) # Add the height
    weight.append(i["attributes"]["weight"]) # Add the weight
d = {'height': height, 'weight': weight} # Put it in a dictionary
df = pd.DataFrame(data=d) # Put it in a dataframe

Step IV: Plot it on a Scatter Plot!

Finally, plotting everything is as simple as two lines:

fig = px.scatter(df, x="height", y="weight", trendline="ols")
fig.show()

And with that, you should be done!

Linear Regression, Image by author

Step V: Explore and Experiment!

That’s it! You’ve now created a linear regression using data from a graph database! Try to query for other variables and run a linear regression using those variables or enhance the graph, adding a title, changing the colours, etc. Good luck!

You can also reference the Colab here: https://colab.research.google.com/drive/1U1mG7gzce3lCLhrsi0KCUZn3z0v-iLca