Are you struggling to get consistent value out of your structured data?
You're not alone! Try this recipe for writing prompts in StoryTell to get the best results.
Part 1. Your Data
Start by describing the data that the model has access. Is it a CSV file with stock data? A text file with customer reviews? Any details about the format and content are helpful.
Part 2. Your Goal
What are you trying to achieve with this request? Do you want to pick a stock to invest in? Do you want to understand Market Sentiment about your product? This is where you leave the model with the broad goal you have in mind.
Part 3. Your Requirements (optional)
How do you want this problem to be approached? Do you want a forecast to be generated? Do you want a table to be created with pros and cons? This is your opporunity to specify details / requirements that should be followed when processing the request.
See it in action:
Attached are 2 documents with data about a specific public company:
- A csv file with daily stock data
- A txt file wite articles about recent company developments
I would like to know if this company would make a sound investment. Please summarize the historic trends in stock price over the last year and then forecast the price 30 days into the future using a SARIMA method.