Simple

At May's R meetup Rich Shepard will present on analyzing censored environmental data using the NADA package.

"Extracting Correct Information From Censored Environmental Data"

Richard B. Shepard, PhD

This will be a casual get together for women who want to hack on projects with or around other women. All types of projects (software and hardware), languages, platforms and experience levels are welcome.

Don't have anything to work on? Come, and we'll help you find a project!

What you should bring:

• Your laptop or other equipment
• Your lovely self!

We'll provide:

• the space, including lots of seating and tables
• internet
• some snacks, including gluten-free and vegan options (Simple's buying pizza for us this time!).

Our goal is to support local women hackers (and aspiring hackers) by providing a safe, welcoming environment in which you can connect with and learn from each other.

We're kid-friendly, too, so feel free to bring your children with you.

Women Who Hack is open to all self-identified women and genderqueer persons, as well as accompanied children of all genders.

All participants of Women Who Hack agree to follow the Citizen Code of Conduct.

Questions? Get in touch: womenwhohack@stumptownsyndicate.org.

Our Google Group is https://groups.google.com/forum/#!forum/women-who-hack.

Until now, nonparametric regression has been a heuristic process - fitting a surface without statistically justifiable confidence measures. With the discovery of the gamma test, there is now a mean squared error estimator that is applicable to the nonparametric case. This allows true statistical reasoning for functional relationships, even when nothing is known about the algebraic form of the relationship. The gamma test is available in R. It will revolutionize the discovery of causal relationships in data.

Wayne Haythorn recently retired from a 30 year career in software design, during which he worked for organizations including NASA, the European Space Agency, National Weather Service, NOAA, Linsoft (Sweden), and Modern Terminals (Hong Kong). He is now working to promote use of the gamma test for nonparametric regression. He is a member of the ASA and has run more gamma tests than anyone else in the world.

Daniel Fennelly will be giving a presentation for January's meetup.

Description:

We'll use the TopicWatchr R package developed to access
Lucky Sort's API to explore several features of time-series text-data.
We'll cover basic text statistics, event detection,
and term clustering for topic modelling.

About the speaker:
Daniel is a software developer at Lucky Sort Inc. in
Portland, OR. He graduated from Reed College in 2011 with a BA in
psychology.

Michael Schilmoeller is presenting on cluster analysis in R for our November meetup! Note that due to proximity with Thanksgiving, this meetup has been scheduled for a Tuesday instead of a Wednesday.

Talk Title:

Cluster Analysis in R

Description:

We will present two applications of the Agnes cluster analysis package. The first will illustrate considerations in selecting a distance metric. The second will show a technique for evaluating the size and number of clusters appropriate for representing data. Both examples are from electric power generation and transmission modeling.

About the speaker:

Michael is a Senior Power Systems Analyst with the Northwest Power and Conservation Council. He holds degrees in physics and business administration, as well as a Ph.D. in mathematics.

Paul deGrandis will be speaking about his experiences building and shipping Clojure and Clojurescript applications.

For this month's meetup, Jack Horne will present on integrating R with Excel in the context of market research.

Jack is interested in the possibility of holding instructional R sessions; let him know if you might like to attend!

This meetup will be at Simple, to whom we extend our thanks for hosting.

Talk Description:

Market research simulators are often delivered using Microsoft Excel. While this software provides an enormous range of functionality, especially when combined with Visual Basic, sometimes it is not enough given either large amounts of data or specific algorithmic needs. The power of R to rapidly process large data sets and to go well beyond the mathematical capabilities of Excel can be leveraged into Excel-based simulators through RExcel.

Speaker Bio:

Jack Horne is a vice president in the Marketing Sciences Group. He has worked as an applied statistician for more than 20 years in fields as diverse as technology and communications; financial services; healthcare and drug development; biochemistry and sensory perception of foods; and consumer packaged goods.

Our August meetup will feature a talk on creating R packages, by Brian Diggs!

Creating a package in R using devtools and roxygen2

One strength of R is its built in support for extending functionality using packages. Packages have specific structure and tools built to enforce some documentation and testing; these lead to a better final product but create a higher hurdle for package creation, especially the first time when the conventions are being learned. This talk will walk through all the steps of creating a package including fulfilling the documentation and testing requirements (and options), resulting in a package that can be distributed and installed by others. A combination of standard tools as well as those provided by the devtools and roxygen2 packages will be used.

At the July meetup, we'll share projects and help one another better use R. Come in with your laptop and get started with R, hack on a personal project, or find advice for coping with your latest R vexation.

This was also the plan for the last meetup, but due to mis-communication about the availability of space at Simple, we weren't able to actually get to the fun part. Apologies to those who attended with high hopes. This month will be a proper realization of last month's plan.

Also, please note that this meeting is on a Wednesday night! This is a departure from recent past meetings, which have occurred on Tuesdays.

Regards,

Homer

Dan from Lucky Sort will tell us about how they use Clojure and their experience using it for serious Java interop. As always, pizza and drinks will be provided by Simple.

Kevin Lynagh will talk to us about data-driven web applications with ClojureScript:

A web page or application is, at its core, just a visual representation of data that people can read, look at, and manipulate. Typically the mapping between the abstract data and elements on the screen is implicit in the code: take some piece of the data, do X here, do Y there; when the user clicks on that thing, modify the page here, and so on.

Reasoning about such code is difficult: either control is inverted across many different callbacks with complected concerns, or one must endure a great deal of ceremony with models, controllers, view models, and views/templates to structure an application. Ideally when we build a web application all we should have to do is describe how our application's data should be represented on the DOM. We shouldn't need to worry about callbacks, twiddling the attributes of particular elements, or updating cached state.

As it turns out, ClojureScript's rich data structures and semantics allows us to easily write such simple, declarative code. I will discuss these ideas using examples from visualization-rich dashboard applications.

Recommended reading:

C2: http://keminglabs.com/c2/
Knockout.js: http://knockoutjs.com/
Rich Hickey's talk on the benefits of immutability: http://www.infoq.com/presentations/Are-We-There-Yet-Rich-Hickey

Soren Macbeth will present on Cascalog and Storm for batch and real-time data processing and machine learning in Clojure.

If you've been meaning to play around with Clojure but never got around to it, you won't want to miss the April meetup of Clojerks. We are going to cover installing Clojure, configuring your editor, using the REPL and building your first Clojure project.

There will be just 1 talk at this meetup. Thanks to Simple for hosting again!

Speaker: De'Mel Mojica

Propensity Score Matching (or “matching” as it’s commonly called) has gained in popularity among methodological circles with its rather parsimonious set of modeling assumptions and relative efficiency.  This talk will briefly introduce users to the basic ideas behind its application; provide an overview of “Matchit” a popular R package that employs matching techniques to broad set data scenarios; and (if time allows) demonstrate its usefulness with a test case.

Alex will go over some of the interesting things from Clojure/West.

There will be 2 talks this March! The R meetup is returning to Simple, to whom we extend our gratitude for hosting. See below for speakers and talk descriptions.

Mike Psaris

This talk will cover some of the functions and packages used to import data into R, and present some of the ways to perform analysis on stream temperatures throughout the Columbia River Basin.  Includes a discussion of some of the packages to incorporate into future work, such as plyr and snow. 

Cameron Bracken

Dynamic report generation is an integral part of statistical analysis.  When you think of dynamic report generation you might think "Sweave" but Sweave is only the beginning of the tools now available in R.  Dynamic report generation in R started with Sweave about 10 years ago but has evolved considerably since then. This talk will cover the origins of Sweave and current extension packages such as cacheSweave, pgfSweave and weaver, as well as other approaches such as brew and the all new and possibly end all knitr package.

A meetup in Portland, OR for people who use the Vim text editor to share tips, knowledge, and various other things they find useful in their daily writing of whatever.

The first meetup will occur in Simple's office where there will be:

• VimScript and why it's not that scary
• other proposals are welcome!
• or I'll just talk about Fugitive.vim and stuff

Please e-mail matt.sacks+vim@simple.com for questions/comments.

Jim Loring and Josh Reich, respectively, will be giving February's talks. See the tentative talk descriptions below. This will be our second meeting Simple; our last one was quite successful!

Thanks to Simple for hosting!

Jim Loring, National Park Service - "Water Quality in R - Revisited"

An example of water quality analysis using R. Using data gathered from Issaquah Creek near Issaquah, Washington and Johnson Creek in Portland, Jim will discuss the context in which the data was gathered, the R algorithm used, an interpretation of the results in the characterization of Issaquah and Johnson Creeks. The presentation illustrates a basic application of R in the management of natural resources and is a continuation of a R Talk given last October.

Josh Reich, Simple - "An Introduction to Machine Learning in R"

Josh will cover k-means, KNN, kernel methods, LDA, rpart & SVMs. A good introduction to techniques geared towards either those new to R or ML, which will draw heavily on material from EOSL.

Josh Reich will be presenting on the data.table package!

Additionally, Simple's Director of Research Allen Goodman will be giving a talk on time series analysis:

"From a statistical perspective, the analysis of time series hasn't dramatically diverted from the program established by George Box and Gwilym Jenkins in 'Time series analysis: Forecasting and Control'. However, from a computational perspective, statisticians haven't adopted best practices for model identification, parameter estimation, model diagnostics, or forecasting. I'll model a time series using the Box-Jenkins program in R, outlining contemporary attitudes within the R community on the subject"

Join us Saturday, June 14th for the 2008 Pearl Pub Pedal, where we will be visiting each of the Pearl District breweries while enjoying special savings on some of the world’s best beer! The event will begin at 1:00 pm at the new Deschutes Brewery located the corner of NW 11th Ave. and Davis St. We will then continue to Bridgeport for $2.75 pints before concluding the tour at Rogue, where pints will be once again offered at reduced prices.

The 2008 Pearl Pub Pedal is proudly sponsored by Pdx Pedicab, UrbanDrinks.com, Neighborhood Notes, and MapClicks.com. Pdx Pedicab will be on location to offer transportation between each of the breweries, making this event a true Portland experience.

If you plan on attending, please RSVP at
http://www.urbandrinks.com/PubPedal