search

blab / nextflu

Real-time tracking of influenza evolution
http://nextflu.org
GNU Affero General Public License v3.0
88 stars 19 forks source link

LBI scaling #85

Closed trvrb closed 7 years ago

trvrb commented 9 years ago

Should LBI be on a strictly relative scale, i.e. between 0 and 1? Other colorings (like dfreq) adjust the color scale so that the dynamic range of the colors are used, but the statistic still has an absolute underlying scale. Would it make sense to move LBI to a similar scheme?

rneher commented 9 years ago

I'll play with this a little. the absolute value depends on few things like the tau and the total number of nodes. but I agree in general it would be better to have this on an absolute scale

trvrb commented 7 years ago

(Mostly) resolved in 141296270a392c176d09f3cb0b2adca8cf586ba3.