First work out what you want to do. If all you want to know is how badly a country has been hit by Covid-19, all you need is the last column in worldometer (the cases per head – this is obviously more important than just the raw number).
If you want to know how badly a country is hit right now, you need to divide the active cases by the population (or perhaps the critical cases by the population). On these measures, China is recovering fast and South Korea is getting there.
However there is a fair chance what you really want to know is how safe a country is to be in, or to admit travellers from. The number of new cases is a starting point for this. If they had perfect detection, and an effective quarantine system, it's all you would need to know. But of course it would mean the detection rate after a few days would drop close to zero (because there would be no transmission). This doesn't happen, so clearly no-one has even near a perfect detection rate. Nevertheless, if all countries had similar detection rates and quarantine effectiveness, it would give a reasonable proxy for ordering countries. Of course that's not true either. How can we tell? If they had similar detection rates, and similar cure rates, then their death rate per head of population would probably converge to a similar figure over time. This isn't happening. Some countries have far too many deaths. One reason is that fatalities are far more likely for older or sicker people. So it's fairer to divide the death rate by the total population over 65, making the figures more comparable. Italy looks worse if you don't do this. Italy also has a high death rate in some areas because the health system is overwhelmed (like Wuhan a month or so ago). So we may underestimate the detection rate (but if the health system is overwhelmed, you probably don't want to go there anyway). Using these proxies, we can tell, for example, that China and South Korea have pretty good detection and quarantine. Confirmation of this comes from the fact that their detection rates are plummeting - they have already quarantined most of the active cases, so spread is winding down. You can also see that Norway and Iceland have pretty terrible per-head infection rates (don't even look at the figures for San Marino, they are terrifying), but they are also on top of it, and their detection rates are also going down. Of course, the gradient of detection rates is also a useful figure (as is the gradient of death rates), but automating this requires retaining data for a reasonable time – say a week – and then matching the tables together. It's not simple.
Finally, some comments on why this may not work for some countries. For starters, I'm assuming that the country is at least able to tell what people died from (that is, when they're close to death, with Covid-19 symptoms, they're going to get tested, and the results will be reasonably correct). That means that the US figures are close to useless for analysis, because they used test kits that didn't work. I don't know how bad the situation is in the US, but it is clearly far worse than the figures show. It also assumes the reported figures are honest. Iran has been accused of faking the statistics. That is incredibly easy to do, if you're prepared to lie, and it's possible to do so undetectably for a while. Eventually, though, it will show up in ordinary death rates. For Iran, their figures so far have looked very self-consistent: but they have plenty of excellent mathematicians, so faking it would be easy. And they have incentives to do so. If the satellite photos of mass graves in Qom are real, then maybe it's starting to show. Unfortunately the provenance of the photos isn't clear (I don't have any time to track it down), and there are plenty of groups with the incentive to mess with Iran, so I just don't know.
Note: I've tried to automate these processes, but tracking down all the macros required to do it has taken more time than I have. If anyone with more expertise in web scripting and backend data analysis is interested, I'd be really happy to work together to build a useful website (needs to use free resources though, I have no money for this).