I'm training a ML model on close price data that is aggregated across 25 exchanges (including dark pools). Eventually I plan to incorporate this model into a pipeline that retrieves price data via the KiteConnect websocket API. I think the best/most efficient way to do this would be via websocket using the last traded price (ltp) mode. On my side I would determine the minutely close price by crosschecking against the current time on my machine. Correct me if you think there's a better approach.
What I'm wondering however is if the data I'm using to train my model is representative of the data I will retrieve via the ltp websocket. i.e is the last traded price aggregated across multiple exchanges? Is there a different endpoint I could use that is?
What exactly does aggregation mean. Are close prices averaged across exchanges, or is the most recent close price across all exchanges used?
You may use Websocket API or Quote API.