Note that it is actually possible to return the predicted values as a hidden attribute of the dataframe, but I think it's less confusing to say to Pingouin's users that if they want a fuller output they should just use as_dataframe=False to get a dictionary instead of a "summary" dataframe
Hi @BikeNW, while this is not directly possible, I believe that you can create additional columns in your dataframe that would represent the contrast that you want to try (e.g. interaction between two terms). Otherwise, I'd recommend checking statsmodels if you'd like to have R-style formulas. Thanks!
hello! first - thank for the fantastic package! i have come across an issue and it may be because of my lack of skills, but cant work it out - it seems that for a mixed anova design you can't have multiple within subject comparisons. e.g. :
pg.mixed_anova(dv='value', within=['variable_a', 'variable_b'], between = 'group', subject='subject', data=df, effsize="ng2") returns an dtype error. is this me being an idiot or is it a limitation? i couldnt find a direct mention of it anywhere so thought id check before deep diving
Hi @macaskill-lab! You're not an idiot: the mixed ANOVA only allows for one between + one within, or two within, or two between factors. It doesn't work with two within + one between, or two between + one within! For more complex ANOVA designs, I'd highly recommend using the free JASP software!
hi all. I'm having some problems getting started with pingouin. I've installed the module via the cmd line and also via conda but when I try to import it in my jupyter notebook it says it cannot find the module, even though it has been successfully installed. any ideas as to why that's the case?
Hi @raphaelvallat , thanks for your reply. I did check and it was the case. In fact, it works now after I restarted my laptop (even though it still didn't work after restarting my jupyter kernel as you have suggested in your message from Jan. 19). So I have no idea how or why but it's all good now :)
dose anyone know how to return a p-value? not the t-test function, i specifically want the p-value
Hi @johnmustin1_twitter, you want the p-value of a T-test? You can get it with this function: https://pingouin-stats.org/generated/pingouin.ttest.html#pingouin.ttest (the
p-val column)
Hi all, I thought you might be interested in knowing that Pingouin has been featured in "towardsdatascience.com"https://towardsdatascience.com/statistical-tests-t-test-andanova-674b242a5274
Hi there! I would like to know if it is possible to tell
pingouin to return the list of outliers found after the robust Shepherd's pi correlation. I checked the source code in master/pingouin/correlation.py, and the function seems to return an array of booleans, indicating whether they are an outlier or not. However, I suppose this function is wrapped in others, so what I get back in the end is the number of outliers in total. It would be useful to find who are those outliers and inspect them further. Is this possible? Thanks for the good work! :-)
Hi @Irene-GM_gitlab ! It is possible by directly calling the lower-level pingouin.correlation.shepherd function (https://github.com/raphaelvallat/pingouin/blob/b369de9b8607e55061bc408e534e7d44443c744a/pingouin/correlation.py#L146-L184):
r, p, outliers = shepherd(x, y)
2 replies
Hello, everyone!
I'm a novice user of Pingouin and I have a question regarding encoding of categorical variables in the
I'm a novice user of Pingouin and I have a question regarding encoding of categorical variables in the
anova method. I noticed that, once I add a column of categorical values (in str format), the code runs without errors, but I wasn't able to figure out what kind of encoding was applied to the data. Could you help me understanding that?
(also, awesome work on Pingouin! I'm having a blast working with it)
Thanks :smiley: ! I haven't tested it extensively with pd.Categorical yet, but generally any kind of groupby operations on the between variable is done with:
grp = data.groupby(between, observed=True) which ensure that categorical levels with no value will be excluded. I would however recommend using string whenever possible!
Hi everyone! I am trying to run a repeated measures anova that looks like this:
aov = mixed_anova(dv='rt', between='experiment', within= ['congruent','PC'], subject='subid', data=df) but I am getting the error: "ValueError: Grouper and axis must be same length". Any ideas? Many thanks!
@xzllxls Pingouin is not tested anymore for Py3.5 so I cannot tell you exactly which function will fail and which will work, but my guess is that most basic functions should world (as long as you're using recent versions of pandas, numpy and scipy). I would however recommend upgrading your Python version to >3.7. Thanks
@raphaelvallat Hi, thanks for your reply. Actually, no data is missing.
@raphaelvallat the data is11790 rows × 2400 columns
@Charlesepfl I do not unfortunately. As you can see in this paper (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3927875/), multivariate normality tests are usually applied on data ranging from 2 to 10 columns (dimensions). Do you really need to test for multivariate normality?
For more details see the
nan_policy argument of the pairwise_corr function https://pingouin-stats.org/generated/pingouin.pairwise_corr.html
Hi, I would like to do the Bonferroni and Sidak multiple comparisons tests following two- and three-way ANOVA. Prism explains these tests at the following link: https://www.graphpad.com/guides/prism/latest/statistics/stat_the_method_of_bonferroni.htm. I have data in two columns (two genotypes) and want to compare them at each row (different levels of my factor). Can Pingouin do these tests?
@DForDespair the VIF is not currently implemented in Pingouin. It can be calculated in statsmodels however: https://www.statsmodels.org/stable/generated/statsmodels.stats.outliers_influence.variance_inflation_factor.html
@csamuel11 Three-way ANOVA are not supported in Pingouin. For 2-way ANOVA, you can use pingouin.pairwise_ttests to calculate the post-hoc pairwise T-test and use the
padjust parameter to specificy no correction / bonferroni / sidak. For more complex ANOVA designs however, I strongly recommend using JASP instead of Pingouin. Hope this helps!
The order of the between subject factor matters for the interaction only. You should choose depending on which interaction order is the most relevant to your data. Alternatively, you can use JASP to calculate all possible pairwise combinations (full interaction as opposed to Pingouin which only reports a "partial" interaction).
Hi @XiaTaopsycho ! Please see: raphaelvallat/pingouin#103 Note that I don't anticipate that LMM will be implemented in the near future. In the meantime, I would recommend using statsmodels or pymer (https://eshinjolly.com/pymer4/)