Hacker News Reader

Comment by waynecochran | Hacker News Reader

Could someone explain the graph used in the article where the y-axis is labeled as "cumulative" -- does this mean that each bin includes the sums from the previous bins? If so, then this will always be a non-decreasing graph. Ebola outbreaks are horrible, but it doesn't do any good to use stupid graphs like this to make it look worse than it is.

View on HN

If they're using the same data source I am at work it's because that is what the ministry of health for the DRC is releasing daily. They don't track active cases but rather cumulative cases in a health zone.

Also if you want to explore the data UN OCHA has it here: https://data.humdata.org/dataset/ebola-cases-and-deaths-drc-...

Yes, you have it exactly right. That graph makes it look like ebola is progressing at a logarithmic rate. It's not. It's nearly linear.

I am all for fighting something as horrible as Ebola, but you can't do this ...

    actual rate    => graph

    const rate     => linear graph
    linear rate    => quadratic graph
    quadratic rate => cubic graph
    ...

WHO reports are much better overall: https://www.who.int/ebola/situation-reports/drc-2018/en/

But you can't trust Nature?

My apologies. I meant to say that the WHO reports are generally more informative and better presented to me. Nature's article is good, but I find the WHO reports better.

No, cumulative is correct if you want to show a rate increase which you clearly see twice based on this graph.

Reply on news.ycombinator.com