Skip to main content
Tutorials

Score prospects in real time using Signals and ML

Solution accelerator

  • Introduction

  • Set up notebook

  • Configure Signals attributes

  • Train the ML prospect scoring model

  • Create intermediary API endpoint

  • See scores in the browser

  • Conclusion

Last updated on

Set up notebook

This tutorial uses a Google Colab Juypter notebook. Follow along in the notebook using your own data.

Start by configuring your credentials in the notebook. Go to Signals > Overview in Snowplow Console to find your Signals credentials. Add them, along with your warehouse and ngrok details, to the notebook secrets:

Here's a list of all the credentials you may need to configure, saved as variables:

from google.colab import userdata

# Snowflake credentials to prepare a training dataset on top of atomic events table
ENV_SF_ACNT = userdata.get('SF_ACNT')   # Account
ENV_SF_USR = userdata.get('SF_USR')     # User
ENV_SF_PWD = userdata.get('SF_PWD')     # Password
ENV_SF_WH = userdata.get('SF_WH')       # Warehouse

# Snowplow Signals credentials
ENV_SP_API_URL = userdata.get('SP_API_URL')       # Signals API URL
ENV_SP_API_KEY = userdata.get('SP_API_KEY')       # Signals API key
ENV_SP_API_KEY_ID = userdata.get('SP_API_KEY_ID') # Signals API key ID
ENV_SP_ORG_ID = userdata.get('SP_ORG_ID')         # Snowplow org ID

# Snowflake Snowplow atomic table
ENV_ATOMIC_TABLE = userdata.get('ATOMIC_TABLE')

# ngrok token to be able to expose the notebook behind HTTPS for querying the API
ENV_NGROK_TOKEN = userdata.get('NGROK_TOKEN')

Adjust these to suit your warehouse authentication methods.

Once you've added the secrets, start working through the tutorial. If you prefer to run the cells in one go with Run all, update your details in the required places first - they're marked with UPDATE THIS.

Last updated on