I am just now reading this and amazed at the broad nature of the responses! Just to throw out a little life science out there - I think the reason the temperatures effect on the final division rates in a much more apparent way is because it is linear in nature while the lux values actually behave in a cyclical way to stimulate the divisions through simulating activity/rest cycles most biological life requires. It may be interesting to analyze or visualize how these cycles effect the growth patterns in some way. Perhaps I will take a look at it later this week.

Keep up the good work guys!

Thank you Vitaliy! And yes are correct, we can do a double-dendrogram plot:

HeatmapPlot[Transpose[Rest /@ sdata[[3 ;; -1]]], 
 Map[Row[{"temp:", Style[#[[1]], Bold, Purple], ", lux:", Style[#[[2]], Bold, Purple]}] &, Transpose[Rest /@ sdata[[1 ;; 2]]]], 
 First /@ sdata[[3 ;; -1]],
 DistanceFunction -> {EuclideanDistance, EuclideanDistance}, 
 Dendrogram -> {True, True}, ImageSize -> 1000]

This plot seems more informative than the first one I posted. Note the grouping of the temperature and lux pairs becomes the same as in the raw data.

Awesome, Anton, thank you! Hierarchical clustering did also crossed my mind, nice execution! From the option

Dendrogram -> {True, False}

I guess that one can add another "orthogonal" dendogram for columns with the option set to {True, True}, is that correct? Similar to this one:

https://www.r-bloggers.com/drawing-heatmaps-in-r

Intraspecies comparison using RadarChart: http://community.wolfram.com/groups/-/m/t/1260507

Beautiful work, thank you for taking the challenge! And thank you very for submitting to Reddit! BTW could "negative value for divisions per day" mean that population is shrinking? Basically death rate > birth rate... assuming they start with some substantial initial population.

Thanks @l van Veen - I opted for a relative change bar chart to show this!

Very Nice. I also like Sanders (and Vitaly) but I noticed with Sanders approach that some algae react opposite to the LUX. With your approach this is not obvious but could be highlighted perhaps. EG. the ellipsoidia is lower with 2500K and higher with 5000K The oculata is the other way around. I don't know if it is really relevant but Sander's approach (less intuitive (for me)) shows this very nicely. Your approach for me is more intuitive. Perhaps a Color code on the top of the bars might indicate this pattern? Just a thought :)

Bubble chart for 4D data: http://community.wolfram.com/groups/-/m/t/1257885