OJ Sales Simulated

This dataset is derived from the Dominick’s OJ dataset and includes extra simulated data to simultaneously train thousands of models on Azure Machine Learning.


Microsoft provides Azure Open Datasets on an “as is” basis. Microsoft makes no warranties, express or implied, guarantees or conditions with respect to your use of the datasets. To the extent permitted under your local law, Microsoft disclaims all liability for any damages or losses, including direct, consequential, special, indirect, incidental or punitive, resulting from your use of the datasets.

This dataset is provided under the original terms that Microsoft received source data. The dataset may include data sourced from Microsoft.

The data contains weekly sales of orange juice over 121 weeks. There are 3,991 stores included and three brands of orange juice per store so that 11,973 models can be trained.

View the original dataset description or download the dataset.


Name Data type Unique Values (sample) Description
Advert int 1 Value indicating if there were advertisements for that orange juice during the week 0: No Advertisements 1: Advertisements
Brand string dominicks tropicana Brand of orange juice
Price double 2.6 2.09 Price of the orange juice (in USD)
Quantity int 10939 11638 Quantity of orange juice sold for that week
Revenue double 38438.4 36036.0 Revenue from orange juice sales for that week (in USD)
Store int 2658 1396 Store number where the orange juice was sold
WeekStarting timestamp 1990-08-09 00:00:00 1992-02-20 00:00:00 Date indicating which week the sales are related to


WeekStarting Store Brand Quantity Advert Price Revenue
10/1/1992 12:00:00 AM 3571 minute.maid 13247 1 2.42 32057.74
10/1/1992 12:00:00 AM 2999 minute.maid 18461 1 2.69 49660.09
10/1/1992 12:00:00 AM 1198 minute.maid 13222 1 2.64 34906.08
10/1/1992 12:00:00 AM 3916 minute.maid 12923 1 2.45 31661.35
10/1/1992 12:00:00 AM 1688 minute.maid 9380 1 2.46 23074.8
10/1/1992 12:00:00 AM 1040 minute.maid 18841 1 2.31 43522.71
10/1/1992 12:00:00 AM 1938 minute.maid 14202 1 2.19 31102.38
10/1/1992 12:00:00 AM 2405 minute.maid 16326 1 2.05 33468.3
10/1/1992 12:00:00 AM 1972 minute.maid 16380 1 2.12 34725.6

Data access

Azure Notebooks

from azureml.core.workspace import Workspace
ws = Workspace.from_config()
datastore = ws.get_default_config()
from azureml.opendatasets import OjSalesSimulated

Read data from Azure Open Datasets

# Create a Data Directory in local path
import os

oj_sales_path = "oj_sales_data"

if not os.path.exists(oj_sales_path):
# Pull all of the data
oj_sales_files = OjSalesSimulated.get_file_dataset()

# or pull a subset of the data
oj_sales_files = OjSalesSimulated.get_file_dataset(num_files=10)
oj_sales_files.download(oj_sales_path, overwrite=True)

Upload the individual datasets to Blob Storage

We upload the data to Blob and will create the FileDataset from this folder of csv files.

target_path = 'oj_sales_data'

datastore.upload(src_dir = oj_sales_path,
                target_path = target_path,
                overwrite = True, 
                show_progress = True)

Create the file dataset

We need to define the path of the data to create the FileDataset.

from azureml.core.dataset import Dataset

ds_name = 'oj_data'
path_on_datastore = datastore.path(target_path + '/')

input_ds = Dataset.File.from_files(path=path_on_datastore, validate=False)

Register the file dataset to the workspace

We want to register the dataset to our workspace so we can call it as an input into our Pipeline for forecasting.

registered_ds = input_ds.register(ws, ds_name, create_new_version=True)
named_ds = registered_ds.as_named_input(ds_name)

Azure Databricks

# This is a package in preview.
# You need to pip install azureml-opendatasets in Databricks cluster. https://docs.microsoft.com/en-us/azure/data-explorer/connect-from-databricks#install-the-python-library-on-your-azure-databricks-cluster
# Download or mount OJ Sales raw files Azure Machine Learning file datasets.
# This works only for Linux based compute. See https://docs.microsoft.com/en-us/azure/machine-learning/service/how-to-create-register-datasets to learn more about datasets.

from azureml.opendatasets import OjSalesSimulated

ojss_file = OjSalesSimulated.get_file_dataset()
# Download files to local storage
import os
import tempfile

mount_point = tempfile.mkdtemp()
ojss_file.download(mount_point, overwrite=True)
# Mount files. Useful when training job will run on a remote compute.
import gzip
import struct
import pandas as pd
import numpy as np

# load compressed OJ Sales Simulated gz files and return numpy arrays
def load_data(filename, label=False):
    with gzip.open(filename) as gz:
        n_items = struct.unpack('>I', gz.read(4))
        if not label:
            n_rows = struct.unpack('>I', gz.read(4))[0]
            n_cols = struct.unpack('>I', gz.read(4))[0]
            res = np.frombuffer(gz.read(n_items[0] * n_rows * n_cols), dtype=np.uint8)
            res = res.reshape(n_items[0], n_rows * n_cols)
            res = np.frombuffer(gz.read(n_items[0]), dtype=np.uint8)
            res = res.reshape(n_items[0], 1)
    return pd.DataFrame(res)
import sys
mount_point = tempfile.mkdtemp()

if sys.platform == 'linux':
  print("start mounting....")
  with ojss_file.mount(mount_point):
    train_images_df = load_data(os.path.join(mount_point, 'train-tabular-oj-ubyte.gz'))

Next steps

View the rest of the datasets in the Open Datasets catalog.