Hi there,
I have a crew completed running off Llama llm.
There is a data analyst and a retailer agent. The purpose is to analyse transaction data and then come up with recommendations for focus products or store issues.
However there is a lot of ‘cleanup’ that needs to happen before the data analyst can work on the data. For example removing specific products that just cause noise (e.g. plastic bags) or ignoring returns etc.
How can I first instruct the data analyst or even create a data cleanup agent to first run through a series of steps to clean the data before it is processed?
Is this a matter of prompts in the task or do I need to pull the data in python first, clean up before I send to agents?
Is this done in the task or agent?