Skipper Seabold

Yeah performance is going to vary, and I don't think you can plan for it, but I would take them at their word. Regions may matter but if you search...

I suppose the lack of consistency only manifests itself in the face of a partition given the CAP theorem, so you really can't plan for it and you won't know...

Semi-relevant https://blogs.aws.amazon.com/bigdata/post/Tx1WL4KR7SE37YY/Ensuring-Consistency-When-Using-Amazon-S3-and-Amazon-Elastic-MapReduce-for-ETL-W Amazon uses DynamoDB as the external system but it itself is eventual consistency AFIAK :)

https://github.com/Netflix/s3mper

You should be able to re-run all models with `results.mle_settings`, if we need to. I found one example in tsa recently where it wasn't correctly attached, but it's fixed now....

Same should apply for WLS.

I used this when I made the regression plots notebook looking at RLM. I think all the R functions automatically handle weights. http://www.ats.ucla.edu/stat/stata/webbooks/reg/chapter4/statareg_self_assessment_answers4.htm

Added a milestone for this. How to use weighted influence statistics for WLS and RLM. Should be low hanging fruit.

Thanks, this is a bug against statsmodels. It looks like you're running the code from github. This code has been updated since the tutorial. You can use get_margeff instead of...

See: https://github.com/jseabold/tutorial/blob/master/preliminaries.py#L18 for the import assumptions. If you prefer I can add them to every file, but I just keep them in a config file.