(Editor's note: Data sources were last confirmed and updated by @BeverlyBrown on 5/22/2024; * denotes most recently added data set.)
Because so many in academia need data for school, I keep an eye out for open sources. Below are sources (alphabetized) that my predecessor Jennifer Scott pulled together and I add to occasionally. Recently, SAS and the National Institutes of Health equipped public health researchers with SAS Analytics Pro on SAS® Viya® to the All of Us Researcher Workbench, the cloud-based platform that allows registered researchers to study data contributed by the All of Us participant cohort. For decades, researchers have relied on SAS to analyze biomedical and clinical data as part of their innovation toolset. Now, researchers will be able to wield their SAS expertise to explore data contributed by more than 400,000 All of Us participants who reflect the rich diversity of the United States, including communities that have been historically underrepresented in biomedical research.
Read on for see the data sets. In the comments, add links to any data sets you've found interesting and useful.
the NIH's All of Us Research Program is building one of the largest biomedical data resources of its kind. Its research hub stores health data from a diverse group of participants from across the U.S. Register for the Researcher Workbench to access data and tools to conduct health research and improve understanding of health and disease.
Project Data Sphere, founded by the CEO Roundtable on Cancer, brings together leading pharmaceutical companies to share data and an analytics platform to investigate those data. The platform now provides access to data from more than 40 studies and 25,000 patients from clinical trials of prostate cancer, breast cancer, and melanoma treatments. The list of available studies is growing rapidly.
Visit this site to tap into a key national resource, serving as a fuel for entrepreneurship and innovation, scientific discovery, and commercial activity.
Obtain public use files to download for analysis, as well as, access pre-digested health statistics.
Where do Citi Bikers ride? When do they ride? How far do they go? Which stations are most popular? What days of the week are most rides taken on? We’ve heard all of these questions and more from you and now we are happy to provide the datasets to help you discover the answers to these questions and more. Developers, engineers, statisticians, artists, academics and other members of the interested public are encouraged to use the data we provide for analysis, development, visualization and whatever else moves you.
Compiled by Deloitte, this is a massive collection of US stats and data sources that, according to the site's "About us," "tells millions of stories about America. Through advanced data analytics and visualization, it tells stories about: places in America—towns, cities and states; occupations, from teachers to welders to web developers; industries--where they are thriving, where they are declining and their interconnectedness to each other; and education and skills, from where is the best place to live if you’re a computer science major to the key skills needed to be an accountant."
Don't get lost in the cool stats and forget to peruse the site's many data sets.
Energy Information Administration
This site offers a number of data sets on energy production, consumption, sources, etc.
Hundreds of data sets on world health, economics, population, etc. All of it is viewable online within Google Docs, and downloadable as spreadsheets.
An international consortium of more than 700 academic institutions and research organizations, ICPSR provides leadership and training in data access, curation, and methods of analysis for the social science research community. ICPSR maintains a data archive of more than 500,000 files of research inthe social sciences.
Kaggle is a site that hosts data mining competitions. Each competition provides a data set that's free for download.
The Million Song Dataset is a freely-available collection of audio features and metadata for a million contemporary popular music tracks.
The Intelligent Systems Division's Prognostics Center of Excellence at NASA hosts a rich data sets you can use to test your problem-solving acumen for a long list of challenges, including managing uncertainty.
Note from @sbxkoenk, who recommended this data source: The data didn't originate with NASA itself. The Prognostics Data Repository is a collection of data sets that have been donated by universities, agencies, or companies. The data repository focuses exclusively on prognostic data sets, i.e., data sets that can be used for the development of prognostic algorithms. Most of these are time-series data from a prior nominal state to a failed state. The collection of data in this repository is an ongoing process.
Scroll down to the "Autophagy networks and databases" heading where you'll find a list of open databases.
Over 1100 Datasets are currently available on the NYC open data portal, more than any other U.S. City
This site helps users get access to research data that is already online or request that data be made accessible. When a user wants access to the data behind a paper, they can make a data request via the app. The app will then see if the data is already available online, and if not, it will contact the author and invite them to make their research data openly available through the Center for Open Science’s Open Science Framework. Once a dataset is delivered, the author will be rewarded with an Open Data Badge to recognise their efforts.
This last one isn't a dataset itself, but rather a social news site devoted to datasets. It's updated regularly with news about newly available datasets.
SAS provides more than 200 data sets in the Sashelp library. 17 of these data sets are used in SAS/STAT documentation. These data sets are available for you to use for examples and for testing code.
Stanford's Large Network Dataset Collection. This list has several datasets related to social networking. Lots of fun in here!
This collection of public data sets, organized topically, covers government, health, space, sports, weather, and much more. There's free data in each set, but there may be a charge for some.
In the UN Comtrade database, you'll find details on countries' import/export habits and much more.
Here you will find data, tools, and resources to conduct research, develop web and mobile applications, design data visualizations, and more.
Dig into real-time data about droughts, earthquakes, floods, landslides, volcanoes, and wildfires.
The U.S. Energy Information Administration (EIA) collects, analyzes, and disseminates independent and impartial energy information to promote sound policymaking, efficient markets, and public understanding of energy and its interaction with the economy and the environment.
Literally hundreds of datasets spanning many decades, sortable by topic or country. Data is downloadable in Excel or XML formats, or you can make API calls. This is an outstanding resource.
Yahoo has various types of data available to share. They are categorized into Ratings, Language, Graph, Advertising and Market Data, Computing Systems and an appendix of other relevant data and resources available via the Yahoo! Developer Network.
The Yahoo News Feed dataset is a collection based on a sample of anonymized user interactions on the news feeds of several Yahoo properties, including the Yahoo homepage, Yahoo News, Yahoo Sports, Yahoo Finance, Yahoo Movies, and Yahoo Real Estate. The dataset stands at a massive ~110B lines (1.5TB bzipped) of user-news item interaction data, collected by recording the user- news item interaction of about 20M users from February 2015 to May 2015.
Some of the CDC data projects, BRFSS at least, can have data in either SAS data set or text files with associated SAS programs to read the data, and associated files with SAS informats and formats. So make sure you get all of the files for the data you are interested in.
Also look for any data dictionaries as there can often be raw data and recoded values and you'll want to know how the recoes were done.
Do not forget data quality checking, there may be oddities in the data especially from surveys.
Thank you SAS for providing these links. 10 years ago people would look for data to work with to learn. This is a starting point for people to get some data. Pistol-stem data set will no longer be the only data set people use.
Thanks shwilliams4 - please let me know if you run across any interesting data sets that we can add to this list.
Jennifer
See also the archived case studies from the Statistical Society of Canada for many interesting (and challenging) datasets from diverse fields at: http://ssc.ca/en/meetings/archived-case-studies
PG
Wow, what a treasure-trove!
Hi Jennifer, I am just a student of SAS learning by my own, by reading Mr. Cody's book named "Learning SAS by examples". book says that data sets and programs can be downloaded from http://support.sas.com/cody. But I could not find it, Can you please help me to look for it. Thanks, Ali
Hi Zuliqar, the link to the data and programs is a bit difficult to find. Here it is: http://support.sas.com/publishing/bbu/zip/62857.zip Comment back here if you're looking for something different!
Outstanding!
Hee hee hee - lots and lots of data for me to play with!! Thank you so much for posting this!!
I came across a blog post containing a list of 20 Big Data Repositories to check out and thought it would be good to add the link here.
http://www.datasciencecentral.com/profiles/blogs/20-big-data-repositories-you-should-check-out-1
Cheers,
Michelle
Wow, @MichelleHomes, this is awesome! Thank you for sharing.
Another resource that may be of interest are the Kaggle Datasets.
This is great, @MichelleHomes! I'll add this to the article above and mention it on the AU board. Got a lead yesterday on a huge stash of Yahoo datasets. We're awash in open data!
So.
Much.
Data.
Mind.
Melting.
🙂
@BeverlyBrown is that a challenge? LOL
Wow, I've often wasted my time looking in vain for suitable datasets to use for motivating marketing analysis case studies in my university teaching, but prospecting through this collection has really got me some great hits in a fairly short time. I'm really impressed and grateful to SAS and the community for curating this. Well Done (SAS) U!
I appreciate that affirmation, @Damien_Mather! Credit my predecessor, @jennifers_sas, who started this stash. Here's a collection I just ran across, with data sets organized topically...government, sports, health, weather, space, etc. Keep using it, we'll keep adding to it!
Hi Free Data Fans!
I realised I had a great repository of links bookmarked (many of them ones people have posted here) and I wanted to add to the collection. I've tried to make sure I haven't duplicated any of the ones already mentioned. Enjoy!
Chris
http://open.canada.ca/en - Canadian government’s open data sets; I’ve been using these for a while and always impressed with the huge variety.
Open Data Toronto – Born and raised in the city of Toronto, Ontario, this is definitely a treasure trove of great stuff. Some aggregated / summarized, some raw, I’ve been using many of these in the Free Data Friday series and in numerous presentations I’ve given.
http://datausa.io - More focused on the visualization of data, but there are some pretty interesting datasets you can get down to. Does take some time to navigate though, and not all have data that can be downloaded.
http://www.uscampgrounds.info/takeit.html - If you want to play with Geographic data, this is a great site; they’ve done a lot of work compiling Campgrounds in North America, and if you are an avid camper, you can easily import the data into your GPS.
https://data.nasa.gov - Another site I used data from for a Free Data Friday (on meteorites!) – a lot of the datasets mean nothing to me, but every so often I find a new one that’s really cool. I’ve also found some pretty cool images, which make great wallpaper!
https://data.gov.uk/data/search - Jumping outside of North America, the UK has done a phenomenal job at compiling a wide range of data sources, in a multitude of formats.
http://opendataforafrica.org - African datasets, including Port statistics, poverty, oil production, and a huge assortment of other topics. The main page has links to different countries that also offer open data, so if you’re looking for a specific region chances are good there will be something useful.
http://www.opensourcesports.com - Hockey, baseball, football, soccer - even cycling, rugby, and swimming! This website has a ton of data on a ton of sports.
http://www.seanlahman.com/baseball-archive/statistics/ - This database is *huge* and allows for exploration into topics like joining multiple tables, not-so-big big data, and time series analysis. I absolutely love this dataset, and use it as often as I can.
Thought the community may like to know about MIT's new visualization tool - a goldmine for data nerds!
@MichelleHomes well there goes my weekend - thanks 🙂
I knew you'd appreciate it @DarthPathos! Have a fun analytical datavizzy weekend...
@MichelleHomes between you and @BeverlyBrown you two are keeping me very happy!! 8-)
"Analytical datavizzy weekend." Epic! @MichelleHomes and @DarthPathos, your comments made me notice the other data sources @DarthPathos posted in this thread in April. Wow...thanks! @Damien_Mather: Did you see that?
Another resourceful list of data for folks - http://flowingdata.com/2016/11/10/find-the-data-you-need-2016-edition/
Yes, thinking of you @DarthPathos and your weekend again
@MichelleHomes LOL I already saw your tweet and already sent it to myself thanks for sharing and hope you have a great weekend!
Chris
@MichelleHomes I appreciate your latest contribution, which benefits anyone looking for practice data and guaranteed to distract @DarthPathos for hours!
@BeverlyBrown @MichelleHomes Hahah - distracting me is definitely not a pro...SQUIRREL!!!!! Oh chocolate!! Oh hey something's shiny over there 🙂
The joys of sharing with the awesome side effect of keeping @DarthPathos occupied! LOL. Have a wonderful weekend guys...
More great teaching data links, thanks team!
I'm loath to post anything negative about free data, especially when we get it for free, but I'd like to reflect some comments from my advanced business analytics class just finished this semester back to the community:
Much of the kaggle data seemed so heavily anonymized to them so as to be unusable for many of their learning and research opportunities.
Maybe this is due in part to the marketing (and therefore customer) orientation of my course, but it seems to me that some datasets are 'overanonymised' if that is indeed a word.
I am keenly aware that useability, as a function of anonymity, for public data is an 'ideal point' (inverted parabolic or 'U' ) relationship, and that without anonymity we would have much less data available, and different domains generally have different views on what constitutes sufficient anonymity, and different protocols to ensure adequate privacy.
Does anyone have a feel for characterizing some of the other teaching data sources on an anonymity scale?
Hey Damien!
Data privacy is a topic that fascinates me (I work in a hospital, so it's critical for what i do). I would love to discuss this and provide my thoughts but I'm exhausted and need to get some work done. Please post back and we'll keep chatting 🙂
Have a great evening
Chris
OK, you asked for it..
Anonymity in New Zealand and Australia for...
Primary research data:
As a university researchers I always promise to store securely, limit access to a small group of named reseachers, analyse and summarise so that individual participants cannot be identified, then eventually destroy after a fixed number of years, all the primary survey and qualitative data I collect. As far as I know, this is mandatory if I want to get ethical approval for my research from my university, and ethical approval for primary research data gathering is in turn mandatory. So nobody outside the prior-nominated researcher group gets to see the data, anonymised or otherwise. Sometimes for removal of potential researcher bias the data is also anonymised for some researchers.
Secondary internal data:
There are commercial customer privacy laws that prevents firms from collecting and retaining customer data unless they can demonstrate that is part of their legitimate business activity, whilst legitimate business activity is understood to include any analysis that is designed to generate improvements in customer value and experience. Generally improvements in customer experience and value can in turn be tied to improvements in business efficiency and profitability. Analysing internal secondary data for anti-competitive and market-dominance-maintaining reasons is explicitly prohibited on penalty of severe fines, as is lax data firewall and handling security. For this type of data to be made publicly available for general learning, an Australasian firm would have to assume responsibility for the effectiveness of any anonymising process, to ensure its customers cannot ever be identified.
So I can sort of understand why corporate legal councel insist on data anonymising protocols that err on the conservative side, i.e. go over the top, before any is released into the wild.
Secondary external data:
This is where it gets a little tricky for me.
Data and metadata that is largely or primarily machine-generated, and does not explicitly identify individual users, like google page stats for selected government public websites, should be OK to analyse and reproduce unless explicitly protected by an agreed protocol, not that any come to mind apart from Creative Commons variations. However I'm never entirely sure that someone smarter than me might be able to identify users somehow.
Any data in the public domain generated by someone typing something, like I am doing right now, is generally protected by author copyright, which means, at a minimum, any explicit reproduction should acknowledge authorship during the period of copyright, and, as in the case of Creative Commons, other protocols should be adhered to. Reproduction and explicit CC or other protocols aside, the data and metadata should be fairly available for analysis, which generally summarises the heck out of what is generally big corpora and structured data and metadata. But when we type, do we ever stop and think how someone might be openly and honestly attempting to analyse what we write, either specifically identifying us, or in bulk with others' creative output? As an academic, I do**, but how widespread is that consciousness? If writers are generally not aware, or limit their awareness to specific domains, how do we, as analysts, ensure their moral rights are fairly protected?
Who analyses the analysts?
If you were tired before reading this, you'll be asleep by now.
Nitey-nite.
** I watch my own bibliographic metadata like a hawk to see who is citing me and to what extent!
Wow, @Damien_Mather that was impressive - and a great way for me to spend my Friday night. I'm sitting here watching Supernatural, eating chocolate ice cream, reading SAS stuff and waiting for my wife to get home ;-).
For me, I have worked for Canada's major telecomm company, 2 hospitals (1 currently) and the Ministry of Health in Ontario. In every job, I have had full access to significantly sensitive material - from customer address, phone bills, and so on to full medical histories (including social work and psychiatric notes). I have taken great pride in my job, and always approach my position that if my data, or that of my parents, were in the database, how would i want it handled? I am highly involved in research at the hospital, and have implemented policies and audits to ensure our data is as protected as possible.
Deidentified data is absolutely critical - but so is data with full PHI, when it's handled appropriately. One of my favourite types of analyses is geospatial cluster analysis - I always manage to find some truly interesting findings when I do this, and the only way I can do it is with postal codes. Bringing age, income etc. into the equation obviously increases the level of identifiable information but could potentially increase the value of the findings exponentially. When i send out reports, I use screen shots to ensure no identifiable information is sent. When I worked at the first hospital, we had a breach because someone sent out a number of pivot tables in Excel without realising that the PHI was still accessible (Excel has hidden sheets with the data, which enables the pivot tables to be manipulated).
Ok - enough rambling from me. All this to say that I have felt for a long time that anyone working with data should be held to a standard (such as the American Statistical Association) where ethics and integrity etc. are clearly outlined, and methods are detailed for handling situations that arise.
Thanks for the food for thought - and here I was thinking I was going to sleep tonight 😉
Chris
I am unable to open the first link (Cancer research data through Project Data Sphere). That one is actually the one I need most. It says "access denied, You do not have sufficient privileges for this resource". Does anyone know how I can access that link or get sufficient privileges? Thanks.
Hi @Primavera, the link should work now. Give it a try!
Awesome. Thanks so much @BeverlyBrown 🙂
Great! Thanks
Are you ready for the spotlight? We're accepting content ideas for SAS Innovate 2025 to be held May 6-9 in Orlando, FL. The call is open until September 16. Read more here about why you should contribute and what is in it for you!
Data Literacy is for all, even absolute beginners. Jump on board with this free e-learning and boost your career prospects.