How to Split Tabular Data for Time-Series Sales Prediction in Vertex AI?

PMLE By islever Editorial Last updated

Short answer

C Use Vertex AI chronological split, and specify the sales timestamp feature as the time variable

Question

You work for a retail company. You have a managed tabular dataset in Vertex AI that contains sales data from three different stores. The dataset includes several features, such as store name and sale timestamp. You want to use the data to train a model that makes sales predictions for a new store that will open soon. You need to split the data between the training, validation, and test sets. What approach should you use to split the data?
  1. Use Vertex AI manual split, using the store name feature to assign one store for each set
  2. Use Vertex AI default data split
  3. Use Vertex AI chronological split, and specify the sales timestamp feature as the time variable Correct Answer
  4. Use Vertex AI random split, assigning 70% of the rows to the training set, 10% to the validation set, and 20% to the test set

Key Takeaways

Tests the application of time-aware data partitioning; the common trap is selecting a random or categorical split that ignores temporal ordering and risks look-ahead bias.

This question evaluates proper data splitting techniques for time-series tabular datasets within Vertex AI. The community consensus strongly favors chronological splitting to prevent data leakage and accurately capture temporal sales trends.

Common mistake: Option A is the most frequently chosen wrong answer because candidates focus on the 'new store' requirement, but grouping by store name fails to account for time-dependent patterns and does not generalize forecasting capabilities to unseen locations.

Expert Analysis

Why the Answer Is Correct

Chronological splitting respects the inherent order of time-series data, ensuring the model learns from historical trends to forecast future values. By using the sales timestamp as the time variable, you prevent look-ahead bias and data leakage, which is critical for retail sales forecasting. Vertex AI’s built-in chronological split automates this process while maintaining strict temporal boundaries across training, validation, and test sets.

Why the Other Options Are Wrong

Random splits (Option D) and default splits (Option B) shuffle data indiscriminately, mixing past and future timestamps, which breaks temporal dependencies and inflates performance metrics artificially. Manual splits based on store names (Option A) group data by category rather than time, failing to simulate how a new store would experience seasonal or trending patterns over time. None of these alternatives address the fundamental requirement of forecasting future sales based on historical temporal progression.

Community Comment Notes

Candidates consistently highlight that time-series data mandates chronological ordering to avoid data leakage, as emphasized in high-voted discussions [Comment 1, Comment 2]. Several users note that random or categorical splits ignore seasonal trends and recent market shifts, making them unsuitable for retail forecasting [Comment 2, Comment 7]. One contributor shared official Google Cloud documentation directly supporting temporal best practices, reinforcing the technical validity of this approach [Comment 4].

Official Reference

https://cloud.google.com/automl-tables/docs/data-best-practices#time

This link opens the official vendor documentation in a new tab.

Exam Strategy

Always identify whether your dataset contains temporal or sequential patterns before selecting a split strategy; if it does, prioritize chronological splitting to prevent data leakage and ensure realistic evaluation. When faced with 'new entity' scenarios, remember that temporal ordering often matters more than categorical grouping for forecasting tasks.

Community Votes

C
85%
A
15%

85% of anonymous learners picked answer C. Votes are pick records left by other test-takers — they are not the verified answer.

Discussion 8 contributions

Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.

From public community threads 8

  1. fitri001 👍 6 Answered C
    Time-Series Data: Your sales data has timestamps, indicating it's time-series data. A chronological split considers the order of the timestamps, ensuring the model is trained on historical trends. Predicting for New Store: Since you want to predict sales for a new store, a chronological split is better than a random split (option D) which wouldn't prioritize recent trends. Vertex AI Functionality: Vertex AI's chronological split functionality is specifically designed for time-series data and leverages the timestamp feature you provide to separate data for training, validation, and testing.
  2. guilhermebutzke 👍 2 Answered C
    My answer C: A: Not Correct: Splitting based on store name wouldn't guarantee temporal separation of data. Furthermore, for this problem is note to assign one store for each set, because the target is for a new store. B: Not Correct: Randomly choosing data points across different time periods could lead to the model not capturing seasonal trends or temporal patterns effectively. C: CORRECT: it leverages the chronological nature of the data. Since the dataset contains sales data over time from different stores, using a chronological split ensures that the model is trained on data from earlier time periods and validated/tested on more recent data. D: Not Correct: Similar to B, a custom random split wouldn't ensure temporal separation and could lead to issues with capturing temporal trends.
  3. pikachu007 👍 1 Answered A
    By using a manual split based on store names, you can train a model that is more sensitive to the unique characteristics of each store, ultimately leading to better predictions for the new store.
  4. Omi_04040 👍 1 Answered A
    Since the question is to predict the for a new store and not sales prediction in general, the answer has to be 'A'
  5. b1a8fae 👍 1 Answered C
    Anything different than option C could potentially lead to data leakage imo.
Show 3 more comments
  1. BlehMaks 👍 1 Answered C
    https://cloud.google.com/automl-tables/docs/data-best-practices#time
  2. vale_76_na_xxx 👍 1
    I say C , time-based splitting is always suggest
  3. shadz10 👍 1 Answered C
    I agree with b1a8fae

Reader comments & corrections

No reader corrections yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

How to Minimize Development Time for Weekly TensorFlow Retraining in Vertex AI?

S-grade

How to Handle Categorical Features in BigQuery ML for Fast Deployment?

S-grade

How to preprocess BigQuery data for TensorFlow?

S-grade

How to Track Model and Prediction Lineage in Vertex AI?

S-grade

Practice All PMLE Questions

Access 65 questions with complete answers and detailed explanations.

View Full PMLE Practice Test →

← Back to PMLE Study Guide