Google BigQuery
Query the warehouse in plain language.
If your reporting data lands in BigQuery, connecting it lets ProhostAI run the queries you would otherwise be waiting on someone to write.
Anteprima
Ecco Google BigQuery nel tuo spazio di lavoro.
Potenzia questi ruoli
Overview
Portfolios large enough to have a warehouse usually have a bottleneck to go with it. The data is complete and well-modelled, and the number of people who can get an answer out of it is small — so questions queue behind whoever writes the SQL.
That queue changes what gets asked. Nobody raises a request for a question they are merely curious about, which means the analysis that gets done is the analysis someone was willing to formally ask for.
Connecting BigQuery removes the queue for the routine half. Ask a question in plain language and get the query run against your own tables, with the operational context ProhostAI already holds sitting alongside it. The warehouse stays exactly as your data team built it.
What ProhostAI can do with Google BigQuery
Once BigQuery is connected, ProhostAI can query your datasets in plain language.
Read
- Run queries from a question Turn a plain-language question into SQL against your own tables. "What was average occupancy by region last quarter?"
- Explore what exists List datasets and tables so you can see what is actually available to ask about.
Create & update
- Manage datasets Create and organise datasets where the connected account has permission.
- Combine with live operations Read warehouse figures alongside what ProhostAI knows about current stays and work.
Setup
BigQuery connects through Google Cloud OAuth against the project holding your datasets.
- Open ProhostAI and go to Settings → Connections.
- Search for Google BigQuery and select Connect.
- Sign in with a Google Cloud account that has access to the project and approve the permissions.
- Ask a question about one of your datasets to confirm the connection works.
Good to know
- BigQuery bills by data scanned. A broad question over large tables can be an expensive question, so connect an account with sensible quotas.
- ProhostAI works within the connected account’s IAM permissions — datasets it cannot read stay unreadable.
- Answers are only as good as the model underneath. Well-named tables and columns produce noticeably better results.
- Consider connecting a read-only service account for analysis rather than one with write access.
FAQ
Can ProhostAI write SQL for me?
Yes. It turns a plain-language question into a query against your own tables and returns the answer, which removes the queue for the routine half of analysis.
Will this run up a BigQuery bill?
It can. BigQuery charges by data scanned, so a broad question over large tables costs real money. Connect an account with sensible quotas.
Does it respect our IAM permissions?
Yes. It works within the connected account’s permissions, so datasets that account cannot read stay unreadable.
Should we connect a read-only account?
For analysis, that is the sensible default — a read-only service account gets the benefit without granting write access to the warehouse.
How do I disconnect BigQuery?
Open Settings → Connections in ProhostAI and disconnect it, or revoke access in your Google Cloud console.