I have a pandas.DataFrame that looks like this:
Group Replicate Time Normed
0 5ng/mLTGFb 1 1 0.924876
1 5ng/mLTGFb 1 2 0.891171
2 5ng/mLTGFb 1 3 0.928782
3 5ng/mLTGFb 1 4 0.842967
4 5ng/mLTGFb 2 1 0.908404
5 5ng/mLTGFb 2 2 0.886511
6 5ng/mLTGFb 2 3 0.911488
7 5ng/mLTGFb 2 4 0.875320
8 5ng/mLTGFb 3 1 0.921771
9 5ng/mLTGFb 3 2 0.887204
10 5ng/mLTGFb 3 3 0.828524
11 5ng/mLTGFb 3 4 0.873509
12 Glyco0.8% 1 1 0.883661
13 Glyco0.8% 1 2 0.823874
14 Glyco0.8% 1 3 0.875522
15 Glyco0.8% 1 4 0.859953
16 Glyco0.8% 2 1 0.880971
17 Glyco0.8% 2 2 0.832171
18 Glyco0.8% 2 3 0.885903
19 Glyco0.8% 2 4 0.859308
20 Glyco0.8% 3 1 0.894803
21 Glyco0.8% 3 2 0.856890
22 Glyco0.8% 3 3 0.862479
23 Glyco0.8% 3 4 0.846343
24 Glyco2.4% 1 1 0.883367
25 Glyco2.4% 1 2 0.859378
26 Glyco2.4% 1 3 0.893720
27 Glyco2.4% 1 4 0.852863
28 Glyco2.4% 2 1 0.871003
29 Glyco2.4% 2 2 0.840247
30 Glyco2.4% 2 3 0.880209
31 Glyco2.4% 2 4 0.849866
32 Glyco2.4% 3 1 0.906474
33 Glyco2.4% 3 2 0.893899
34 Glyco2.4% 3 3 0.861791
35 Glyco2.4% 3 4 0.841760
36 Glyco4.8% 1 1 0.909556
37 Glyco4.8% 1 2 0.867863
38 Glyco4.8% 1 3 0.875383
39 Glyco4.8% 1 4 0.849258
40 Glyco4.8% 2 1 0.912481
41 Glyco4.8% 2 2 0.862332
42 Glyco4.8% 2 3 0.898971
43 Glyco4.8% 2 4 0.851969
44 Glyco4.8% 3 1 0.920611
45 Glyco4.8% 3 2 0.858697
46 Glyco4.8% 3 3 0.877669
47 Glyco4.8% 3 4 0.842820
48 NegativeControl 1 1 0.896742
49 NegativeControl 1 2 0.828705
50 NegativeControl 1 3 0.898466
51 NegativeControl 1 4 0.840793
52 NegativeControl 2 1 0.889273
53 NegativeControl 2 2 0.830354
54 NegativeControl 2 3 0.838096
55 NegativeControl 2 4 0.843467
56 NegativeControl 3 1 0.859525
57 NegativeControl 3 2 0.823689
58 NegativeControl 3 3 0.833555
59 NegativeControl 3 4 0.840500
This data can be plotted quite nicely with seaborn and matplotlib (assuming the above data was assigned to a variable called normed)
import matplotlib.pyplot as plt
import seaborn as sns
sns.barplot(x='Time',y='Normed',data=normed,hue='Group')
plt.legend(loc=(1,0.3))
Here is the output:
However, what I really want to do is perform a t-test to test for statistical significance between groups (ideally one for all n choose 2 combinations of variable but for now lets just say between the negative control at the first and second time point). Does anybody know if seaborn supports this or know of another way of performing t-tests on this data? Additionally I've seen that seaborn supports bootstrapping statistics, it would be great if I could get confidence intervals for the t-tests as well though this is peripheral really.
Thanks for the help
