John Ramey

On the [Kubernetes blog](http://blog.kubernetes.io/2016/07/thousand-instances-of-cassandra-using-kubernetes-pet-set.html), there's a good example of using `PetSets` to deploy a 1000-node Cassandra cluster.

- [x] Removed `LoadBalancer` from `kafka-cluster.yaml` - [ ] Test [automatic kubernetes DNS name](http://kubernetes.io/docs/admin/namespaces/) with `kafkacat` using a basic job - [ ] Update `README`

I don't have anything to contribute, but thanks @discdiver for adding a ticket so quickly based on my Slack message!

@cdeterman Thanks for asking. Yes, the package is still maintained -- I have not spent any time making updates as of late. I'm glad you pointed out that the CRAN...

There's been some movement on this front. See [these slides](http://urbaninstitute.github.io/R-Trainings/accesing-census-apis/presentation/index.html) for details.

Slides refer to the [censusapi](https://github.com/hrecht/censusapi) package.

Similar to #26, a speedup can be had by replacing `base::merge`. The obvious replacement is to use `dplyr::left_join`, but the latter does not produce the same exact same outcome. Moreover,...

The above was wasted yak shaving. Instead, I realized that no merge step was needed. Moreover, the code was simplified by updating the code for `fips_colors` to: ``` r fips_colors...

The time to render is now due to two sources. 1. The filter time 2. The render time The filter requires approximately 800 milliseconds and is due to: ``` r...

It turns out Dan was right. I was using `base::merge` to join some manually defined state information with the populations obtained from the US Census Bureau. The join was flawed....