How to best normalize results from load testing based on data received?

I’ve run a variety of load tests on my API and the time of the http_req_duration result is proportional to how much data the request returned, which makes sense. Is there a way I can normalize the data so I can say that given this much data we can expect the http_req_duration to take X ms? Should I do data_received / http_req_duration? Any help would be appreciated!

I am not sure I understand exactly what you want to do, but take a look at the various ways you can work with the metrics data in k6:

handleSummary() might be enough if you don’t care about every single request, otherwise something like the JSON or CSV or InfluxDB output might be better.