Hadley Wickham in Portland!
Sounds amazing:
Category: News and Announcements
jobs for people who can work in R
For those who may be interested:
No R Club Week 9
Heads up: No R club next week (but feel free to gather without your facilitators and express your nerdliness).
natural language processing in R
http://www.districtdatalabs.com/#!natural-language-processing-with-r/cv0k/linkedin
It’s expensive, but it looks so cool. Sounds like it includes some basic r-savviness stuff (setting up projects in R studio, etc.), regular expressions, and then some specialized NLP packages. Swoon!
End of Term Reminder
Just a friendly reminder, since we’re approaching the end of the term. If you’re taking R Club for credit, you need to:
1) attend R Club (we’ve been keeping track of attendance, so you’re covered there)
2) send a one-page-ish write-up to Sanjay (sanjay@uoregon.edu) by the end of the term, describing what you’ve done and/or what you’ve learned in R Club this term.
If you are signed up for credit and you think you’ll have trouble meeting one or both of these requirements, email Sanjay.
What’s lined up for R Club?
Next week (week 7) is Regular Expressions Show-and-Tell! John, Matt, and Rose will have some short presentations on how they’re using regular expressions in R (and/or python, in Matt’s case), and then we can spend the rest of the time asking questions, maybe working in small groups if folks have some more targeted questions for any of the presenters.
Week 8 will be a consultation for Matt. He started describing his project a little bit, and so far it sounds like “a big ball of wibbly wobbly timey wimey… stuff,” to quote The Doctor.
Don’t miss it! It’s going to be awesome.
What’s lined up for R club?
Next week we’ll have another consultation about analyzing breaks in time series data (i.e. when you have a long string of repeated measures, how can you tell when a participant “breaks” from his/her pattern?). We’d love to consider a range of approaches, from basic time series models like ARIMA through more flexible multilevel modeling approaches (which may actually make the most sense, because the data are nested). To see what time series stuff we cover in 612, see this rpubs document (the second half covers autocorrelation and times series). If you want to get in the mood, go to Naomi’s brown bag talk on Friday (tomorrow) 12pm-1pm in Straub 256. She’ll be talking about an MLM approach for these data.
Week 6 will be map day! Yes, maps, like for finding treasure. Except these maps won’t be about treasure so much as outreach and recruitment… There are so many great mapping packages in R! I did a little mapping stuff last year for a class in library sciences, so I jotted down some code real quick to give you an idea of how some of this could work:
# let's say we posted flyers at prince pucklers, moss street day care, and peace health hospital
# and let's say two families told us they saw flyers at moss street
library(ggmap)
library(ggplot2)
flyer1 = geocode("moss street day care eugene or") # this is literally a google maps search!
flyer2 = geocode("1605 E 19th Ave Eugene, OR 97403") # this is the address for prince pucklers
flyer3 = geocode("peace health hospital eugene or")
family1 = flyer1
family2 = flyer1
map.data = rbind(flyer1, flyer2, flyer3, family1, family2)
map.data$type = c(rep("flyer", 3), rep("family", 2))
map.data
base.map = qmap("eugene, or", zoom=14, maptype="roadmap") # get a map of eugene (again, this is just a google map search)
map = base.map +
geom_point(data = map.data, aes(x = lon, y = lat, color=type, shape=type), size=8, alpha=0.5) # add the flyer and family info
qmap("university of oregon", zoom=14, maptype="watercolor") # just because :)
And for week 7, we’re going to talk about model selection in machine learning/statistical learning/data mining/whatever you want to call it. Into the data mines!
Longitudinal data deserves longitudinal r club, no?
Thanks to those who have already posted handy code snippets from our meeting this week, working on Jocelyn’s data! If you have something you generated during the meeting that you think might be useful to others, please comment on the original post or Jacob’s data manipulation post to share the code. (If you’re using rstudio, publishing to rpubs can be an easy way to share code without having to worry about wordpress mangling the formatting.)
We did some fun data management, data reduction, and exploratory work, but as far as I know, no one actually got to the point of running any SEM stuff. How to you all feel about continuing with this project next week as well? Now that I’ve got my head wrapped around the design and data, I’d love to start playing with lavaan, which is one of my very favorite toys. On the horizon, we also have some neat time series stuff waiting, and several lovely nuggets (including a “highlights from 612” post to add to the highlights from 611 posts), so I don’t want to spend too much time on this project at the expense of other topics. But since we have 10 meetings this term instead of 5, it seems okay to spend two days on SEM. Especially since the data are longitudinal…
End of Term Reminder
Just a friendly reminder, since we’re approaching the end of the term. If you’re taking R Club for credit, you need to:
1) attend R Club (we’ve been keeping track of attendance, so you’re covered there)
2) send a one-page-ish write-up to Sanjay (sanjay@uoregon.edu) by the end of the term, describing what you’ve done and/or what you’ve learned in R Club this term.
If you are signed up for credit and you think you’ll have trouble meeting one or both of these requirements, email Sanjay.
R Club & | Bayes Club
Play choose your own adventure with our very own data science club line-up! You come to a fork in the road. Do you take it? If so, turn to page round(runif(1,1,100*pi)).
Seriously, though: we have pretty much no days left to make this decision, so please respond to the below as soon as you can. It’s only a few questions long. And thank you so much for taking the time to do so — it’ll be really helpful!