This content originally appeared on DEV Community and was authored by nk_Enuke
Overview
The goal is to build a prototype that processes unstructured data held by companies (chat logs, PDFs, screenshots, recordings, etc.) using BigQuery's AI capabilities to solve real-world business problems.
I'm currently studying GCP data engineering and thought this would be perfect, so I decided to participate.
The content is unique in that you select from 3 approaches (minimum 1 selection):
- Utilizing BigQuery's generative AI features
- Vector search: ML.GENERATE_EMBEDDING, VECTOR_SEARCH, etc.
- Integrated analysis of structured/unstructured data: Object Tables, ObjectRef, Multimodal DataFrame, etc. Also, the evaluation criteria are different from typical Kaggle competitions:
- Technical Implementation (35%): Code quality, effective use of BigQuery AI
- Innovation and Creativity (25%): Novelty of solution, business impact
- Demo and Presentation (20%): Clarity of problem definition, documentation quality
- Assets (20%): Quality of blog/video, public GitHub repository
- Bonus (10%): Feedback, survey submission
Competition Without Dataset
The most distinctive feature is that no data is provided
Participants are free to choose and use publicly available datasets...
For example, the following datasets are recommended:
BigQuery public datasets overview page
BigQuery public dataset marketplace
Image datasets provided through Cloud Storage
Code Reading
Code by Dao Sy Duy Minh
https://www.kaggle.com/code/daosyduyminh/simple-tutorial-weather-forecasting
Demonstrates 3 approaches for weather forecasting using BigQuery.
Appears to be using Hanoi weather data.
- Python and Prophet It seems you can use Prophet through bigquery.Client().
- BigQuery ML You can retrieve datasets with dataset = bigquery.Dataset(dataset_id) and use them directly. Usage is as simple as writing client.query(train_query), requiring relatively little code.
- BigQuery Generative AI Large-scale predictions are possible using FROM AI.FORECAST
This content originally appeared on DEV Community and was authored by nk_Enuke
nk_Enuke | Sciencx (2025-08-19T15:00:00+00:00) BigQuery AI – Building the Future of Data:Day1. Retrieved from https://www.scien.cx/2025/08/19/bigquery-ai-building-the-future-of-dataday1/
Please log in to upload a file.
There are no updates yet.
Click the Upload button above to add an update.