Quickstart: KumoRFM

Coming soon: KumoRFM as an NVIDIA NIM container

This documentation is still being updated ahead of the public release. Please stay tuned.

In this guide, we walk you through an end-to-end example using a multi-table E-Commerce dataset. You’ll learn how to load data, auto-infer a relational graph, and run instant predictions: demand forecasting, churn prediction, item recommendations, and missing value imputation, all without training a model.

Step 1: Install the SDK

Install the Python SDK

1 pip install kumoai

Step 2: Get an API key

You need an API key to make calls to KumoRFM. Use the API key and API URL provided for your KumoRFM environment.

Step 3: Initialize a client

Use your API key to initialize a Kumo client:

1 import kumoai.rfm as rfm, os
2 
3 os.environ["KUMO_API_KEY"] = "ENTER_YOUR_API_KEY_HERE"
4 
5 rfm.init()

Step 4: Import your data

We will use an E-Commerce dataset for this example:

1 import pandas as pd
2 
3 dataset_url = "s3://kumo-sdk-public/rfm-datasets/online-shopping"
4 
5 users_df = pd.read_parquet(f"{dataset_url}/users.parquet")
6 items_df = pd.read_parquet(f"{dataset_url}/items.parquet")
7 orders_df = pd.read_parquet(f"{dataset_url}/orders.parquet")

Step 5: Create a graph

Auto infer links between tables to form a graph

1 graph = rfm.Graph.from_data({
2     "users": users_df,
3     "items": items_df,
4     "orders": orders_df,
5 })
6 
7 # Inspect the graph - requires graphviz to be installed
8 graph.visualize()

Step 6: Make a prediction

Write a predictive query to generate instant predictions:

1 model = rfm.KumoRFM(graph)
2 
3 # Forecast 30-day product demand
4 query1 = "PREDICT SUM(orders.price, 0, 30, days) FOR items.item_id=1"
5 result1 = model.predict(query1)
6 display(result1)
7 
8 # Predict customer churn
9 query2 = "PREDICT COUNT(orders.*, 0, 90, days)=0 FOR users.user_id IN (42, 123)"
10 result2 = model.predict(query2)
11 display(result2)
12 
13 # Item recommendation
14 query3 = "PREDICT LIST_DISTINCT(orders.item_id, 0, 30, days) RANK TOP 10 FOR users.user_id=123"
15 result3 = model.predict(query3)
16 display(result3)
17 
18 # Missing value imputation
19 query4 = "PREDICT users.age FOR users.user_id=8"
20 result4 = model.predict(query4)
21 display(result4)