# Compute reduced summary statistics
**URL:** https://discourse.pymc.io/t/compute-reduced-summary-statistics/4971
**Category:** Questions
**Created:** [April 29, 2020, 10:48am UTC](https://discourse.pymc.io/t/compute-reduced-summary-statistics/4971 "2020-04-29T10:48:49Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@CorbinFoucart](https://discourse.pymc.io/u/CorbinFoucart)
#### Post date: [April 29, 2020, 10:48am UTC](https://discourse.pymc.io/t/compute-reduced-summary-statistics/4971/1 "2020-04-29T10:48:49Z")
Normally I compute summaries of traces with `pm.summary(trace)` but this can be quite slow for large problems. Is there an argument to specify that I only want mean and standard deviation?
[https://docs.pymc.io/api/stats.html?highlight=pymc3%20summary#pymc3.stats.summary](https://docs.pymc.io/api/stats.html?highlight=pymc3%20summary#pymc3.stats.summary) specificies that I can pass a func\_dict of functions, but I don’t see a `pm.stats.mean` function, for example. Does PyMC3 use `np.mean` under the hood?
---
### Author:  [@nkaimcaudle](https://discourse.pymc.io/u/nkaimcaudle)
#### Post date: [April 29, 2020, 11:56pm UTC](https://discourse.pymc.io/t/compute-reduced-summary-statistics/4971/2 "2020-04-29T23:56:57Z")
You can reduce the number of variables that pm.summary() calculates all the stats for by using the var\_names parameter. As for only computing a few of the stats I don’t think that is possible. To check model convergence you should calculate and check rhat and ess at least once after sampling.
Alternatively you can write a simple function to calc just mean and std using numpy and return a DataFrame.
---
### Author:  [@OriolAbril](https://discourse.pymc.io/u/OriolAbril)
#### Post date: [April 30, 2020, 8:36am UTC](https://discourse.pymc.io/t/compute-reduced-summary-statistics/4971/3 "2020-04-30T08:36:33Z")
It looks like you want something like the second example in the documentation, that is, passing a custom `stats_funcs` dictionary and using `extend=False`. Example code below:
```auto
stats_funcs = {"mean": np.mean, "std": np.std}
az.summary(idata, stats_funcs=stats_funcs, extend=False, kind="stats")
```
I’d also like to point out a couple of extra details to make the answer more complete.
The first comment is that PyMC3 now uses ArviZ for statistics and plotting, you’ll see that I have used `idata` instead of `trace`. ArviZ uses its own data structure (aiming to be backend agnostic and therefore be used not only by PyMC3 users but also by PyStan, Pyro…). If performace is critical, you should convert your trace to InferenceData using [`az.from_pymc3`](https://arviz-devs.github.io/arviz/generated/arviz.from_pymc3.html#arviz.from_pymc3), otherwise, the conversion is done internally every time you call an ArviZ function like `az.summary(trace,...)`.
The second comment is that ArviZ uses `xarray.apply_ufunc` with a wrapper around the functions in `stats_funcs`. This wrapper allows users to use a wider range of functions (otherwise it would be limited to functions following numpy ufuncs conventions) but for functions which are already pure ufuncs, it may not be as efficient as doing something like `idata.posterior.mean(dim=("chain", "draw"))`. By default, `az.summary` uses this second approach, but to customize the output, the wrapper must be used. This being said, unless the model has a relatively large number of variables, this should not affect performance significantly.
And finally, I have used the `kind` argument, present only in ArviZ development version to avoid the computation of diagnostics (ess, rhat, mcse). This is bugish, even though the values for the diagnostics are not returned when `extend=False`, [they are still executed](https://github.com/arviz-devs/arviz/blob/master/arviz/stats/stats.py#L1163), unlike diagnostics, [default stats functions are not executed if `exclude=False`](https://github.com/arviz-devs/arviz/blob/master/arviz/stats/stats.py#L1112).
---
### Author:  [@CorbinFoucart](https://discourse.pymc.io/u/CorbinFoucart)
#### Post date: [April 30, 2020, 12:56pm UTC](https://discourse.pymc.io/t/compute-reduced-summary-statistics/4971/4 "2020-04-30T12:56:30Z")
Thank you – that’s incredibly helpful. Marking as solution.