Basically, I have two 7x7 tables, which are saved as long_iccs and pvalues in data.
For the long_iccs table, the x-axis is titled as 03m, 06m, 09m, 12m, 18m, 24m, and 36m, and the y-axis is also titled as 03m, 06m, 09m, 12m, 18m, 24m, and 36m-- I want to keep this format in my output graph.
03m | 06m | 09m | 12m | 18m | 24m | 36m | |
---|---|---|---|---|---|---|---|
03m | 1 | 0.25 | -0.002 | 0.10 | -0.057 | -0.063 | NaN |
06m | 0.25 | 1 | 0.384 | 0.265 | 0.209 | 0.021 | 0.209 |
09m | -0.0021 | 0.374 | 1 | 0.454 | 0.426 | -0.0132 | 0.361 |
12m | 0.100 | 0.265 | 0.454 | 1 | 0.286 | 0.118 | 0.414 |
18m | -0.0572 | 0.209 | 0.426 | 0.286 | 1 | 0.248 | 0.409 |
24m | -0.0632 | 0.0213 | -0.0132 | 0.118 | 0.248 | 1 | -0.163 |
36m | NaN | 0.209 | 0.361 | 0.414 | 0.409 | -0.163 | 1 |
For the pvalues dataframe, it looks like this:
V1 | V2 | V3 | V4 | V5 | V6 | V7 | |
---|---|---|---|---|---|---|---|
1 | NaN | 0.146821255 | 0.5032691552 | 0.2744200984 | 0.696125633 | 0.68544985 | 0.809989049 |
2 | 1.468213e-01 | NaN | 0.0046248855 | 0.0255566484 | 0.084404798 | 0.42822440 | 0.081061255 |
3 | 5.032692e-01 | 0.004624886 | NaN | 0.0004159696 | 0.002517379 | 0.53866956 | 0.006688649 |
4 | 2.744201e-01 | 0.025556648 | 0.0004159696 | NaN | 0.040601011 | 0.21443715 | 0.002414332 |
5 | 6.961256e-01 | 0.084404798 | 0.0025173792 | 0.0406010109 | NaN | 0.07311703 | 0.008312505 |
6 | 6.854499e-01 | 0.428224404 | 0.5386695582 | 0.2144371492 | 0.073117029 | NaN | 0.840898661 |
7 | 8.099890e-01 | 0.081061255 | 0.0066886493 | 0.0024143322 | 0.008312505 | 0.84089866 | NaN |
Now talking about what the output graph should look like:
-
basically it should look like heatmap (these values are from the long_iccs):
-
then, on each block, I want to add *, **, *** at the top right corner of the numbers if the corresponding values from the pvalues table are <.05, <.005, < .0005 respectively (if the block shows NaN, no need to add any asterisks).
I have tried to "hack" the corrplot package but the significance level given to me is not based on my pvalues table, but the correlations of long_iccs. Some brilliant people please help, thanks a lot!
https://stackoverflow.com/questions/67376993/how-to-make-a-heatmap-like-graph-with-significance-level-based-on-two-tables-in May 04, 2021 at 07:32AM
没有评论:
发表评论