He has "interpolated" missing data points as lying on the graph between the existing points. But we do have data points, in the raw data. What we see is that, in terms of the "per 100,000" column H if the value of dead was 0,1 or 2 the data was marked as a -. Meaning "to low to meaningfully calculate". If you wanted to graph it the correct value would be ZERO. Not what ever value is between the last two numbers.
The graph is wildly incorrect. (I figured this out doing averages of the two data sets being graphed) Which means that the assertion being made using the graph is also wildly incorrect.
Full disclosure: I am not in favor of vaccinations. I am not vaccinated and have no plan to do so.
Having said that, the headline is misleading at best. Table 6 in the data shows that 92,711 unvaccinated people died and 229,981 vaccinated people died. Table 7 shows that the study included 4,363,715 unvaccinated people and 30,885,740 vaccinated people. According to those numbers the likelihood of dying is:
Unvaccinated: 92,711 / 4,363,715 = 2.12% died.
Vaccinated: 229,981 / 30,885,740 = 0.74% died.
So using the numbers from the study, the likelihood of unvaccinated people dying is 3 times higher than for those who are vaccinated according to this study.
So the data in this study does NOT support saying that “Vaccinated people under 60 are twice as likely to die as unvaccinated people”. This data shows quite the opposite.