Then forget to do this extraction, as you will get a false estimate of HR (I tried this myself a couple of times). The only way is to email the authors and kindly ask for their data, or for univariate HR and CI. Although it is rare to get any response, sometimes it works (try to find the email of the first author, as corresponding authors are usually busy).

There is no safe method that you can use to extract this information. Please don't do this :) If you play with an example, you will see unreliability of extracting HR without having the number at risk

You can email the corresponding author though, and ask for more data

@@am1rsafavi-naini356 Yah I think no one will reply, unfortunately, but I have been able to benefit a lot from your both videos.

Sorry, I am afraid I can't. If you are a student, I guess doing it on your own will pay for the time you spend on it. But anyway, good luck.

You can find the expected events in the manuscript text usually (for example the total number of survived, or percent of events, ... in that group). Remember that papers that didn't present events at each time point are usually low-quality, and this method of estimating HR is susceptible to error.

Yup, email me at [email protected] and we will arrange a meeting to do one of them