I've analyzed monks, phenotypes, and colleges electoral
I know the homes of Pompeii, and I hum the tunes historical
I find and classify your friends, with labels categorical.

I'm very well acquainted, too, with e-mail topic tracking,
I understand reductions, both from min cut and set packing ,
I love my inference algor'ms, even when they're hella hard,
like when I'm taking products with operators Hadamard.

I'm very good at authorship and finding out graphology,
I can tell your rank from meetings, your web history, psychology.
In all application areas, central, dull, or peripheral
I am a topic modeler of data that are general.

I know our mythic history, Landauer and Tom Ha-hoffman,
my code is open source, for I've released it all on C-R-A-N
I make users click on buttons, that lead to pretty portraiture
At every stock I look to predict fiduciary investiture
I can tell split spectrographs from cepstral plots and excel files
I know the reason why you think your code refuses to compile
Then summarize a rag of which I've read the writer's gist afore,
And tell you who wrote that infernal nonsense Pinafore.

Then I can speed up sampling schemes for MCMC inference
then tell you which of many tags have just a tad of coherence
In all application areas, central, dull, or peripheral
I am a topic modeler of data that are general.

For my publication record, though its lengthy and cite-wo-orthy,
only gives you nu-um-bers in terms of the perplexity
In all application areas, central, dull or peripheral
I am a topic modeler of data that are general.