AI-Powered Variable Search
The search feature described in this article is powered by AI, but is entirely self-contained and does not use generative AI or connect to any large language models (LLMs). It is designed to enhance search accuracy and relevance through advanced algorithms, without the need for external models or real-time generation of content. This ensures that all search results are based on the specific dataset and context, providing a focused and reliable experience.
You can use Crunch’s AI-powered search to find your variables. AI-powered search allows you to use natural language to search for data, which means you can type your search results the same way you would normally speak.
Searching for a variable
- In a dataset, locate the search bar on the top left of the page:
- Type your search criteria in the box and then click Enter. Remember, this is an AI-powered search so you can use natural language to find what you’re looking for.
- Your search results appear. The most relevant search results will appear first.
- Click or drag & drop a variable in the search results to analyze the data in Crunch.
Providing feedback on your search results
- Click the Provide feedback on search results button.
- A feedback form appears:
- Rate your search results and provide other relevant feedback.
- Click Submit when finished.
Indexing datasets
All data must be indexed before you can use the AI-powered search. You can use the following information if you find a dataset that hasn’t been indexed yet.
If your dataset hasn’t been indexed yet, the following message appears below the AI search bar:
Click Index dataset to index the dataset. This should take a few minutes, but large datasets may take longer. Once the dataset has been indexed, the message is removed and you can begin using AI-powered search on this dataset.