Analytics – Page 43 – DECISION STATS

SAS makes it free to learn SAS

Wanna learn SAS language? Why go elsewhere? Get the original software in a virtual machine from the guys who actually created it!

http://www.sas.com/en_us/offers/14q1/122603-sas-for-academia/overview.html

It’s a 2 1.4 GB download and does require a 2 minute registration but thats all. You can even protype in it before you take that all important call to go for which software to buy for your analytics need. Note – retraining existing SAS analysts in R is expensive in terms of training costs, business disruption and code migration. Now only if someone could create a software to change SAS code into R code – that would make things different !!

SAS language analysts continue to be in much more demand than R language analysts- partly due to the failure of an authentic official R certification as it does not exist ( a fact that has been exploited by various institutes that teach you language syntax but not the analytics) and the poor state of the mothership R website with its huge and sometimes over detailed documentation

https://communities.sas.com/community/sas-analytics-u

and I really like the professionalism as well as the generosity of the institute. Someone famous once predicted SAS will disappear by 2015 ( based on a time series model ). With billions of dollars and it’s kitty and decade long shrewd and street smart management, SAS is going to stay up a wee bit longer and give it’s wannabe killers a bloody nose or two.

Thank you Professor Goodnight

—

testing it myself now !!

—

from

Click to access sas-university-edition-107140.pdf

For whom is SAS® University Edition designed?
SAS University Edition is designed for anyone wanting easy access to statistical software to perform quantitative analysis in an academic setting. This includes undergraduate and graduate students, professors, adult learners, high school students, researchers and postdocs

DOWNLOAD URL

http://www.sas.com/en_us/software/university-edition/download.html

Screenshot

KDNuggets Poll- Which software do you use to crunch?

The KDNuggets is an awesome poll for measuring which software you did use in the last 12 months. It already has 10000 voters already. What is unique about this poll is –

takes less than 50 seconds to fill
uses verification (single click) to ensure no duplicate votes are done
transparent publishing of results in real time
independent of vendor conflict of interests and honest.
you can even download the raw data and start crunching your own analysis

I have many friends who send me emails to vote for them, and I have voted. Plus I wrote occasionally for the site.

So here is the link for the powerful KDNuggets.com poll –http://www.kdnuggets.com/

http://vote.sparklit.com/poll.spark/203792

FOAS is now going to reduce your tax if you support open source

Announcement from the retiring founder of http://www.foastat.org/

—

FOAS is a nonprofit public benefit corporation registered in California. It has federal tax-exempt status under Internal Revenue Section 501(c)(3). FOAS has a worldwide mission to to promote free software, open access publishing, and reproducible research in statistics.

FOAS now has 501(c)(3) federal tax exempt status. All donations to FOAS, retroactively from September 2012, are now tax deductible.

—

What is FOAS- http://www.foastat.org/

The mission of the Foundation for Open Access Statistics (FOAS) is to promote free software, open access publishing, and reproducible research in statistics.

FOAS works to ensure the continued success of the Journal of Statistical Software (JSS), one of the few major open access journals that is free for both readers and authors. We also promote the use and development of free software for statistics, such as the R language and environment for computational statistics. We encourage members and the academic community at large to publish reproducible research that is publicly available online, e.g., in an open access journal or on an open access pre-print server.

You can join FOAS to show your support for free statistical software, open access publishing, and reproducible research in statistics. Membership is free and open to all. Members receive periodic email updates from FOAS, and may wish to join the FOAS-members list, join the Facebook FOAS Group, or follow FOAS on Diaspora.

Consider making a donation to FOAS. All donations are currently being used to support the operation of JSS.

—-

eh, Google how about routing less money through the Irish and giving some to Californian open source…

Related- cultural icon to non open source companies.

This is why code written by scientists gets ugly

brilliant piece of why scientists write bad code

What You're Doing Is Rather Desperate

There’s a lot of discussion around why code written by self-taught “scientist programmers” rarely follows what a trained computer scientist would consider “best practice”. Here’s a recent post on the topic.

One answer: we begin with exploratory data analysis and never get around to cleaning it up.

An example. For some reason, a researcher (let’s call him “Bob”) becomes interested in a particular dataset in the GEO database. So Bob opens the R console and use the GEOquery package to grab the data:

Bob is interested in the covariates and metadata associated with the experiment, which he can access using pData().

Bob discovers that pd$characteristics_ch1.2 is “age at examination”, but it’s stored as a factor. He’d like to use it as a numeric variable. So he sets about figuring out how to do the conversion.

Three levels of nested methods. Ugly. However, it works, so Bob moves to…

View original post 34 more words

The Data Scientist Song

Dining in San Francisco – Let R Guide You

Rolling Your Rs

I’m frequently asked by newcomers to R to provide an easy to follow generic set of instructions on how to download data, transform it, aggregate it, make graphs, and write it all up for publication in a high impact journal – all by the end of the day ! While such a request is somewhat understandable coming from a student it’s somewhat awkward when coming from a seasoned research investigator, (i.e. someone who should know better). Obviously, no such set of instructions exists given the astonishing variety of available data stored in differing formats with varying degrees of adherence to a standard(s). Simply put, “real world” data wrangling and analysis can be a complex process independent of the language you choose to employ. Nonetheless, what I’ll attempt to do here is provide a “tour” involving some data I picked at random. I’ll be sticking to the problems of downloading, filtering…

View original post 1,413 more words

Analyzing Facebook Networks using #rstats

Analyzing Facebook Networks using #rstats is given below. Note this can be replicated by anyone- and you can theoretically be able to create adjacency matrix (friend of friends) of all people who have given permission into your app- thus maybe a network analysis of people to some group (like R Facebook group https://www.facebook.com/groups/rusers/?fref=ts or even FOASStat https://www.facebook.com/groups/foastat/?fref=ts is possible

While traditional marketing and targeting focusses on variables like demographics and psychographics, social network analysis lets you tap into influential people within communities with the right offer and then let the free retweet, reshare as proxy free ads. The hypothesis being that people in similar network clusters have similar consumer behaviours (not just in similar demographic or psychographics). Also this should ideally be able to capture changing consumer behaviour faster for better targeting.

Sources-

1) Go to

https://developers.facebook.com/apps

2) Create a new app
https://developers.facebook.com/apps/241634542707615

3) Change url in settings ( to http://localhost:1410 ). I wish we could do this in twitteR package and oAuth for TWitter but I have no clue . that. Also coming up post on Google Plus and LinkedIn (if I can)

https://developers.facebook.com/apps/241634542707615/settings/

3) Copy the secret thingies from the dashboard. DO click the screenshots to enlarge (i.e to understand what I am trying to do here)

https://developers.facebook.com/apps/241634542707615/dashboard/

4) Modify code below with the secret thingies above

install.packages("Rfacebook")
install.packages("httpuv")
library(Rfacebook)
library(httpuv)
#fbOAuth(app_id, app_secret, extended_permissions = TRUE)
fb_oauth=fbOAuth("241634542707615", "4f2e6273c08d1b56d1bf14626d9e19a6", extended_permissions = TRUE)
#fb_oauth <- fbOAuth(app_id="123456789", app_secret="1A2B3C4D")
save(fb_oauth, file="fb_oauth")
getwd()
dir()

Created by Pretty R at inside-R.org

5) Hey now we have a fb ouath object we can just load it up. we the following code can then be modified at will to get your analysis

library(Rfacebook)
library(httpuv)
library(RColorBrewer)
 
load("fb_oauth")
me <- getUsers("me", token=fb_oauth)
#getFQL(query, token)
my_friends <- getFriends(token=fb_oauth, simplify=F)
str(my_friends)
table(my_friends$relationship_status)
 
pie(table(my_friends$relationship_status),col=brewer.pal(5, "Set1"))
 
table(my_friends$location)
pie(table(my_friends$location),col=brewer.pal(20, "Greens"))
pie(table(my_friends$locale),col=brewer.pal(4, "Blues"))
pie(table(my_friends$gender),col=brewer.pal(3, "Oranges"))
 
load("fb_oauth")
 
mat <- getNetwork(token=fb_oauth, format="adj.matrix")
 
library(igraph)
 
network <- graph.adjacency(mat, mode="undirected")
getwd()
setwd("C:/Users/dell/Desktop")
 
pdf("network_plot.pdf")
 
plot(network ,vertex.size=5, 
     vertex.label=NA, 
     vertex.label.cex=0.45,
     edge.arrow.size=1,
     edge.curved=TRUE,)
 
dev.off()

Created by Pretty R at inside-R.org

6) Voila my FB graph

and analyzed for how are my homies

Europeans into open relationships more

Places my friends are

Languages they speak

and all with R

Note- this is what an adjancency matrix is- just a fancy word for turning your Fb friends into a 1-0 matrix based on their relationships. The more friends you have the more interesting this analysis

Now lets do likes analysis

I use this code and my own username (since I am the only one who has allowed the app permissions!)

likes=getLikes(user="byebyebyer",n=500,token=fb_oauth)

Boy I like a lot of websites huh- I wonder if I can cluster/ data reduce/association analysis them (but that’s coming up or is it?)

Hat tips-

Analyze Facebook with R

http://blog.revolutionanalytics.com/2013/11/how-to-analyze-you-facebook-friends-network-with-r.html

Please share:

Please share:

Please share:

Please share:

Please share:

Please share:

Please share: