Here is a wonderful example of a geeky nerdy corporate player encouraging education in the liberal arts ( the designers of the GUIs and the phones) of the future.
Google sponsored Doodle 4 Google. (also quite a challenge to traditional brand managers who want to so control the image of the brand- I once waited 12 days for an official Logo to appear on this blog)
HOW TO ENTER: To qualify for entry, go to the sasCommunity.org web site located at http://www.sascommunity.org/wiki/Main_Page
between April 11, 2011 and May 9, 2011 and either add or edit valid content as described herein to earn award points.
Creation of a first time profile on www.sascommunity.org will earn 1,000 points. For each valid article creation or edit, 100
points will be earned. Articles and subsequent edits should adhere to the sasCommunity.org terms of use as outlined on
http://www.sascommunity.org/wiki/sasCommunity:Terms_of_Use. All points’ accumulation will end at 5:00 PM GMT on
May 9, 2011 and only those points earned between 8:00 AM GMT on April 11, 2011 and 5:00 PM GMT on May 9, 2011
will be counted in this contest. Contest entries made through the Internet will be declared made by the registered user of
the sasCommunity.org profile account. Sponsor is not responsible for phone, technical, network, electronic, computer
hardware or software failures of any kind, misdirected, incomplete, garbled or delayed transmissions. Sponsor will not be
responsible for incorrect or inaccurate entry information, whether caused by entrants or by any of the equipment or
programming associated with or utilized in the contest.
ELIGIBILITY: The contest is open to all sasCommunity.org members 18 year of age or older on the start date of the
contest. Void where prohibited by law. Employees (including immediate family members and/or those living in the same household of each), the Sponsor, members of the sasCommunity.org Advisory Board, SAS Global Users Group Executive Board, their advertising, promotion and production agencies, the affiliated companies of each, and the immediate family members of each are not eligible.
PRIZE: Three (3) prizes will be awarded based on total points accumulated during the contest as follows:
1stPlace: 3 SAS®Press books - not to exceed $250 in combined retail value;
2ndPlace: 2 SAS®Press books - not to exceed $150 in combined retail value; and
3rdPlace: 1 SAS®Press book - not to exceed $100 in retail value.
Contribute content or SAS code to sasCommunity.org for your chance to WIN! To qualify, simply add or edit articles between April 11, 2011 and May 9, 2011 (GMT). Creation of a first-time profile on sasCommunity.org gives you 1,000 points. For each valid article creation or edit, 100 points will be earned. The user with the most points collected during this time wins SAS Press Books!
Contributing and gaining points also gets you closer to sasCommunity Guru status as explained in the article below:
Become a sasCommunity Guru
The sasCommunity support team has been hard at work adding new features and is pleased to announce a points system that recognizes each user’s contributions to the site. Every time you contribute by creating a page, updating it, or just doing a little wiki gardening, you earn points.Earning points is automatic and simple – all you have to do is contribute! Creating your account starts you with 1000 points and all the current users have been credited with points dating back to the site coming online in April 2007.
More than 71 Million individuals in the United States are admitted to
hospitals each year, according to the latest survey from the American
Hospital Association. Studies have concluded that in 2006 well over
$30 billion was spent on unnecessary hospital admissions. Each of
these unnecessary admissions took away one hospital bed from someone
else who needed it more.
The goal of the prize is to develop a predictive algorithm that can identify patients who will be admitted to the hospital within the next year, using historical claims data.
Official registration will open in 2011, after the launch of the prize. At that time, pre-registered teams will be notified to officially register for the competition. Teams must consent to be bound by final competition rules.
Registered teams will develop and test their algorithms. The winning algorithm will be able to predict patients at risk for an unplanned hospital admission with a high rate of accuracy. The first team to reach the accuracy threshold will have their algorithms confirmed by a judging panel. If confirmed, a winner will be declared.
The competition is expected to run for approximately two years. Registration will be open throughout the competition.
Data Sets
Registered teams will be granted access to two separate datasets of de-identified patient claims data for developing and testing algorithms: a training dataset and a quiz/test dataset. The datasets will be comprised of de-identified patient data. The datasets will include:
Outpatient encounter data
Hospitalization encounter data
Medication dispensing claims data, including medications
Outpatient laboratory data, including test outcome values
The data for each de-identified patient will be organized into two sections: “Historical Data” and “Admission Data.” Historical Data will represent three years of past claims data. This section of the dataset will be used to predict if that patient is going to be admitted during the Admission Data period. Admission Data represents previous claims data and will contain whether or not a hospital admission occurred for that patient; it will be a binary flag.
The training dataset includes several thousand anonymized patients and will be made available, securely and in full, to any registered team for the purpose of developing effective screening algorithms.
The quiz/test dataset is a smaller set of anonymized patients. Teams will only receive the Historical Data section of these datasets and the two datasets will be mixed together so that teams will not be aware of which de-identified patients are in which set. Teams will make predictions based on these data sets and submit their predictions to HPN through the official Heritage Health Prize web site. HPN will use the Quiz Dataset for the initial assessment of the Team’s algorithms. HPN will evaluate and report back scores to the teams through the prize website’s leader board.
Scores from the final Test Dataset will not be made available to teams until the accuracy thresholds are passed. The test dataset will be used in the final judging and results will be kept hidden. These scores are used to preserve the integrity of scoring and to help validate the predictive algorithms.
Teams can begin developing and testing their algorithms as soon as they are registered and ready. Teams will log onto the official Heritage Health Prize website and submit their predictions online. Comparisons will be run automatically and team accuracy scores will be posted on the leader board. This score will be only on a portion of the predictions submitted (the Quiz Dataset), the additional results will be kept back (the Test Dataset).
Once a team successfully scores above the accuracy thresholds on the online testing (quiz dataset), final judging will occur. There will be three parts to this judging. First, the judges will confirm that the potential winning team’s algorithm accurately predicts patient admissions in the Test Dataset (again, above the thresholds for accuracy).
Next, the judging panel will confirm that the algorithm does not identify patients and use external data sources to derive its predictions. Lastly, the panel will confirm that the team’s algorithm is authentic and derives its predictive power from the datasets, not from hand-coding results to improve scores. If the algorithm meets these three criteria, it will be declared the winner.
Failure to meet any one of these three parts will disqualify the team and the contest will continue. The judges reserve the right to award second and third place prizes if deemed applicable.
I recently found an interesting example of a website that both makes a lot of money and yet is much more efficient than any free or non profit. It is called ECOSIA
If you see a website that wants to balance administrative costs plus have a transparent way to make the world better- this is a great example.
World’s largest tropical forest reserve (38,867 square kilometers, or about the size of Switzerland)
Home to about 14% of all amphibian species and roughly 54% of all bird species in the Amazon – not to mention large populations of at least eight threatened species, including the jaguar
Includes part of the Guiana Shield containing 25% of world’s remaining tropical rainforests – 80 to 90% of which are still pristine
Holds the last major unpolluted water reserves in the Neotropics, containing approximately 20% of all of the Earth’s water
One of the last tropical regions on Earth vastly unaltered by humans
Significant contributor to climatic regulation via heat absorption and carbon storage
Click per milli (or CPM) gives you a very low low conversion compared to contacting ad sponsor directly.
But its a great data experiment-
as you can monitor which companies are likely to be advertised on your site (assume google knows more about their algols than you will)
which formats -banner or text or flash have what kind of conversion rates
what are the expected pay off rates from various keywords or companies (like business intelligence software, predictive analytics software and statistical computing software are similar but have different expected returns (if you remember your eco class)
NOW- Based on above data, you know whats your minimum baseline to expect from a private advertiser than a public, crowd sourced search engine one (like Google or Bing)
Lets say if you have 100000 views monthly. and assume one out of 1000 page views will lead to a click. Say the advertiser will pay you 1 $ for every 1 click (=1000 impressions)
Then your expected revenue is $100.But if your clicks are priced at 2.5$ for every click , and your click through rate is now 3 out of 1000 impressions- (both very moderate increases that can done by basic placement optimization of ad type, graphics etc)-your new revenue is 750$.
Be a good Samaritan- you decide to share some of this with your audience -like 4 Amazon books per month ( or I free Amazon book per week)- That gives you a cost of 200$, and leaves you with some 550$.
Wait! it doesnt end there- Adam Smith‘s invisible hand moves on .
You say hmm let me put 100 $ for an annual paper writing contest of $1000, donate $200 to one laptop per child ( or to Amazon rain forests or to Haiti etc etc etc), pay $100 to your upgraded server hosting, and put 350$ in online advertising. say $200 for search engines and $150 for Facebook.
Woah!
Month 1 would should see more people visiting you for the first time. If you have a good return rate (returning visitors as a %, and low bounce rate (visits less than 5 secs)- your traffic should see atleast a 20% jump in new arrivals and 5-10 % in long term arrivals. Ignoring bounces- within three months you will have one of the following
1) An interesting case study on statistics on online and social media advertising, tangible motivations for increasing community response , and some good data for study
2) hopefully better cost management of your server expenses
3)very hopefully a positive cash flow
you could even set a percentage and share the monthly (or annually is better actions) to your readers and advertisers.
go ahead- change the world!
the key paradigms here are sharing your traffic and revenue openly to everyone
donating to a suitable cause
helping increase awareness of the suitable cause
basing fixed percentages rather than absolute numbers to ensure your site and cause are sustained for years.
Ubuntu has a slight glitch plus workaround for installing the RCurl package on which the Google Prediction API is dependent- you need to first install this Ubuntu package for RCurl to install libcurl4-gnutls-dev
Once you install that using Synaptic,
Simply start R
2) Install Packages rjson and Rcurl using install.packages and choosing CRAN
6) Uploading data to Google Storage using the GUI (rather than gs util)
Just go to https://sandbox.google.com/storage/
and thats the Google Storage manager
Notes on Training Data-
Use a csv file
The first column is the score column (like 1,0 or prediction score)
There are no headers- so delete headers from data file and move the dependent variable to the first column (Note I used data from the kaggle contest for R package recommendation at
Once you type in the basic syntax, the first time it will ask for your Google Credentials (email and password)
It then starts showing you time elapsed for training.
Now you can disconnect and go off (actually I got disconnected by accident before coming back in a say 5 minutes so this is the part where I think this is what happened is why it happened, dont blame me, test it for yourself) –
and when you come back (hopefully before token expires) you can see status of your request (see below)
> library(rjson)
> library(RCurl)
Loading required package: bitops
> library(googlepredictionapi)
> my.model <- PredictionApiTrain(data="gs://numtraindata/training_data")
The request for training has sent, now trying to check if training is completed
Training on numtraindata/training_data: time:2.09 seconds
Training on numtraindata/training_data: time:7.00 seconds
7)
Note I changed the format from the URL where my data is located- simply go to your Google Storage Manager and right click on the file name for link address ( https://sandbox.google.com/storage/numtraindata/training_data.csv)
to gs://numtraindata/training_data (that kind of helps in any syntax error)
## Load googlepredictionapi and dependent libraries
library(rjson)
library(RCurl)
library(googlepredictionapi)
## Make a training call to the Prediction API against data in the Google Storage.
## Replace MYBUCKET and MYDATA with your data.
my.model <- PredictionApiTrain(data="gs://MYBUCKET/MYDATA")
## Alternatively, make a training call against training data stored locally as a CSV file.
## Replace MYPATH and MYFILE with your data.
my.model <- PredictionApiTrain(data="MYPATH/MYFILE.csv")
At the time of writing my data was still getting trained, so I will keep you posted on what happens.
How many accounts in Facebook are one unique customer?
Does 500 million human beings as Facebook customers sound too many duplicates? (and how much more can you get if you get the Chinese market- FB is semi censored there)
Is Facebook response rate on ads statistically same as response rates on websites or response rates on emails or response rates on spam?
Why is my Facebook account (which apparently) I am free to download one big huge 130 mb file, not chunks of small files I can download.
Why cant Facebook use URL shorteners for the links of Photos (ever seen those tiny fonted big big urls below each photo)
How come Facebook use so much R (including making the jjplot package) but wont sponsor a summer of code contest (unlike Google)-100 million for schools and 2 blog posts for R? and how much money for putting e education content and games on Facebook.
Will Facebook ever create an-in house game? Did Google put money in Zynga (FB’s top game partner) because it likes
games 🙂 ? How dependent is FB on Zynga anyways?
So many questions———————————————————— so little time
The training dataset includes several thousand anonymized patients and will be made available, securely and in full, to any registered team for the purpose of developing effective screening algorithms.
DECISION STATS 