Influenza: Number of Sequences over time graphs for multiple clades

[anna-parker]

I saw these lovely graphs in https://www.science.org/doi/10.1126/science.adq0072#abstract, they give a very clear and concise overview of the proportion of a specific variant/clade of all sequences at a point in time. I think that these graphs are easier to understand than our current Number of Sequences over Time graphs and extend better to multiple clades - I could see them being immediately useful for influenza. However, we might have to limit the number of clades/variants that can viewed at a specific time.

[chaoran-chen]

Yes, agreed, that would be very useful! I wonder, should this be a new component or an extension of the prevalence-over-time?

If we extend prevalence-over-time, we could add an optional field like stratifyByField. The bar chart would then be stacked as shown in your example and this would also work if more than one dataset is provided. But I'm not sure how the line and bubble chart would work with multiple datasets (see storybook) – but maybe be can start with disabling the line and bubble charts for now if both stratifyByField and multiple datasets are provided?

[anna-parker]

I was thinking we could extend the Number of Sequences over Time graph then we do not need to think about how to modify the confidence intervals that are in prevalence over time line graph, prevalence is always a ratio to the total and here the bars should add up to a total so it is more Number of Sequences over Time in my understanding

[chaoran-chen]

Ah, yes, sounds good, we can start with Number of Sequences over Time which should be easier. The second row of the plot that you posted shows the proportions and I can see use cases for both but starting with the simpler version makes sense to me.