1 of 37

Free and Open Data�Where to find it and Can you use it

Business Reference and Services Section (BRASS)

of Reference User Services Association (RUSA)

in American Library Association (ALA)

March 18, 2020: Data in Libraries Webinar Series

Presenter: Jennifer C. Boettcher, Georgetown University

2 of 37

Jennifer C. Boettcher

Jennifer C. Boettcher and Leonard M. Gains. Industry Research Using the Economic Census. Greenwood Press: Phoenix, AZ. 2004

M.B.A., Georgetown University, Washington, D.C., 2005

M.L.S., State University of New York, Albany, N.Y.,1992

B.A., University of New Hampshire, Durham, N.H., 1987

ALA RUSA BRASS Member since 1991

Georgetown University 1997-present

Founder of Business Information Finders (BIF) and Capital Area Business Academic Librarians (CABAL) in DC

2013 Emerald Research Grant: Zombie List (reanimated business sources)

Seeking contributors: https://boettcher.georgetown.domains/HisBusColl

2010 Gale Cengage Learning Award for Excellence in Business Librarianship

3 of 37

Librarian & Information Scientist

As a Librarian, I

Understand the source
Know how to find the source
Know the related subjects
Know how it’s connected to other sources
Know how to read it
Make connections between publisher and researcher

As a Librarian, I don’t

Publish the primary source
Have your context or expertise
Do statistical analysis
Interpret the data
Do data entry
Have legal expertise

These are my views and do not reflect those of Georgetown or RUSA.

Boettcher, J. C., & Dames, K. M. (2018). Government data as intellectual property:

Is public domain the same as open access? Online Searcher, 42(4), 42-48.

4 of 37

Data Vocabulary

Datasets: Raw or statistical numbers, can be flat file as Comma Separated Variable (CSV) or Proprietary like Excel (see Bobray Bordelon’s RUSA Presentation, March 4)
Metadata: Variables or fields in the record (example, Author)
Big data: Transactional (example, each check out)
Visualization and Reports: Making sense of what the data means. How many checkout, usually as aggregated statistics (See Ryan Womack, March 25)
Application program interface (API): piece of software that interfaces between getting the data out and putting in your computer (See Jeremy D. Darrington, April 15)

Open Data: Freely accessible data, created for a specific purpose

5 of 37

Adaptations of DIKW pyramid by US Army Knowledge Managers,

from https://en.wikipedia.org/wiki/DIKW_pyramid

Data are not:

Information

Technology

Digital

Analytics

Evidence

Research

Visualizations

Ideas

Data are

collected facts

“raw material”

6 of 37

Copyright provides the owner of copyright with the exclusive right to

Display the work
Distribution of Public Domain collections called “collective works” or compilations
Copyright also provides the owner of copyright the right to authorize others to exercise these exclusive rights, subject to certain statutory limitations.

Reproduce the work in copies
Prepare derivative works based upon the work
Distribute copies of the work to the public by sale or other transfer of ownership or by rental, lease, or lending
Perform the work publicly live or by means of a digital transmission

7 of 37

Copyright and Numeric Data

Facts are not copyrighted (In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work.) 17 USC 102b

In US collections of facts or data that fail to meet the minimum threshold of creativity also are ineligible for copyright protection, even if assembling such a collection takes significant time, effort, or resources: “sweat of the brow.”

Creative expression of data in compilation is protected (Feist 1991)

8 of 37

Public Domain: No Copyright Restrictions

Public Domain is not protected by intellectual property laws, like copyright. Anyone can use a public domain work without obtaining permission, but no one can ever own it.

Example: no longer protected due to age of creative work.

Works produced for the U.S. Government by its officers and employees should not be subject to copyright. The provision applies the principle equally to unpublished and published works. 17 USC 105

REMEMBER: Public domain data must be attributed.

9 of 37

Data policy in the Federal Government

Federal government policy

Passed by Congress
Implemented by Executive Branch
Refined by Courts

Policy on data collection (priorities)

Mandated by law (in CFR) Public Law No: 115-435 (signed 1/14/19)
Implemented by regulations (Federal Register) 82 FR 52213
Directed by memorandum (Presidential) M-13-13
Standard of practice: Data plans (Agencies)

10 of 37

Caveats of Open Government Data�

Why not?

Classified
Not widely distributed/web
Contracted or grant work
Lack of supposed interest, DoJ
Lack of funding, Census
Not kept, Record Retention

Free
Transparent
Accountable
Accessible to citizens
Engage all citizens
Machine Readable, API v Human
When in doubt openness prevails

11 of 37

Why Open Data exists

Funded research created for a specific purpose

US National and some state data
Some other Countries
Non-Government Organizations (NGO)
Grants (e.g. PubMed Central is now “PMC”)
Institutional and public repositories and Open Access publishing

Open Access is not Intellectual Property law it’s a license agreement from the copyright owner and a set principles: CC0

12 of 37

Public Domain Vs. Open Access

Public Domain

US Law
Federal Government products
Data at any stage can be retrieved by FOIA
Some sub-nationals
Some countries
Some NGOs

Open Access

Decision of publisher/owner
Because of ownership of copyright
License (CC, GNU, etc)
Principles

Reuse and redistribution of the data
Allows derivative works as Open only
No restrictions on who can access and use
Electronically transferable
Machine-readable

13 of 37

Data as input and output

Input

FREE data

Public Domain
Open Data

Questionable status

Internet
Repositories

Commercial (licensed) sources

Grants, co-authors, permission

Output

Working with scholars

FAIR

Findable: The first step in (re)using data is to find them. Metadata and data should be easy to find for both humans and computers.

Accessible: Once the user finds the required data, she/he needs to know how can they be accessed, possibly including authentication and authorisation.

Interoperable: The data usually need to be integrated with other data.

Reusable: The ultimate goal of FAIR is to optimise the reuse of data.

14 of 37

Questions?

CC0, https://pixabay.com/en/hedgehog-child-young-hedgehog-1759027

15 of 37

Major Sources of Social Science Data in the US Government

16 of 37

Major Sources of Natural Science Data from the US Government

https://www.flickr.com/photos/notbrucelee/6897137283/in/photostream

17 of 37

P.E.S.T. Analysis for Industry

Political

Legislative

Congress.gov

Executive

Regulations.com

Judicial

United States Courts

Economic

Sector Inflation

BLS’s Producers Price Index

Microeconomic trends

Socio-cultural

Norms & Ratios

IRS’s Statistics of Income

Peers and partners

Technology

Patents

Citation Analysis

Tech Transfer

Government funded research to license

18 of 37

Problems that come with government data

Beggars can’t be choosers

Too old
Not to the geographic level needed
Too detailed
Have to file a FOIA request

Compatibility

Standardization
Combining two datasets even from same source might not be possible
Combining two different sources must look at methodology

19 of 37

Major International Data Sources

By topic

http://data.un.org

Social & Economic- World Bank

Financial & Economic- International Monetary Fund

Human body- World Health Org

Labor- International Labour Org

Telecommunications- International Telecommunications Union

Governance- Transparency International

Developed Countries- Organisation for Economic Co-operation and Development (OECD)

By Country

National Statistical Offices

More data available in national language

Some charge for access

Citizens of that country might have free access

National Repositories/Archives

Historical

Datasets

20 of 37

Where to start

New Jersey https://data.nj.gov

United States https://www.data.gov

International http://data.un.org

Inter-university Consortium for Political and Social Research repository https://www.icpsr.umich.edu/icpsrweb/ICPSR/

21 of 37

Where to learn MORE

For Librarians

For Federal data

22 of 37

Let’s discuss

boettcher@georgetown.edu

202 687-7495

Twitter: @jenny.wombat

These slides are licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

AMSTAT images from

http://magazine.amstat.org/blog/2018/05/01/fy18fedbudget

23 of 37

Numeric Data: who is responsible

Analytics

Analysts

Knowledge
Context of numbers

Tacit Knowledge

Expert

Wisdom
“I Just know it”

Curation

Repository

Preserving access

Datum (single of data)

Researchers

Data
Creation of numbers

Statistics

Publishers

Information
Relations among numbers

24 of 37

Numeric Data: who is responsible

Analytics

Analysts

Knowledge
Context of numbers

Tacit Knowledge

Expert

Wisdom
“I Just know it”

Curation

Repository

Preserving access

Datum (single of data)

Researchers

Data
Creation of numbers

Statistics

Publishers

Information
Relations among numbers

25 of 37

Vocabulary: Tools, Process, and Products

Datasets or compilation: Raw or statistical numbers, can be flat file such as Comma Separated Variable (CSV) or proprietary like Excel

Metadata: Includes field descriptions for the dataset, found in codebooks

Schema: How data is organized or structured using standards, like classification

Application Program Interface (API): Read-only machine to machine querying, generally from JSON or XML files

Big data: Raw, unstructured data; normally transactional (example: each check out)

Natural Language Processing (NLP): Use for text analysis, not numeric data

Artificial Intelligence (AI): Includes predictive analytics and machine learning

Reports: Usually aggregated statistics based on big data (example: how many checkouts)

Data Visualization: Using software to visually communicate relationships and context of data

Open Data: Freely accessible data, created for a specific purpose; by-product of decision making or research

26 of 37

F.A.I.R data

Findable: The first step in (re)using data is to find them. Metadata and data should be easy to find for both humans and computers.

Accessible: Once the user finds the required data, she/he needs to know how can they be accessed, possibly including authentication and authorisation.

Interoperable: The data usually need to be integrated with other data.

Reusable: The ultimate goal of FAIR is to optimise the reuse of data.

Mainly for scientific literature and in Europe

GoFAIR

27 of 37

Works produced for the U.S. Government:�Lifecycle of Data

Policy Makers who ask the questions about what has to be found or measured

Researchers who design methods or experiments to collect the data and where the data and codebooks are created.

Statisticians who manipulate datasets using models and algorithms to see trends in longitudinal data and to interpret data at a moment of time in cross-sectional studies.

Analysts who see patterns using predictive analytics, seek the emerging relationships between the numbers, transforming data into information by giving it context.

Other Data Scientists will link graphics, statistical downloads, and application programming interfaces (APIs) to the researcher's raw data.

Writers and Data Visualization Designers, who uses their imagination to apply their knowledge to make data understandable in reports, press releases, and other resources.

The federal agency will act as Publishers, putting the synthesized resources on its website for all-primarily for decision makers but also for citizens-to read.

28 of 37

Funding for Federal Data Collection

NIH- National Institutes of Health (HHS)

NSF- National Science Foundation

AHRQ- Agency for Healthcare Research & Quality (HHS)

FDA- Food & Drug Agency (HHS)

BEA- Bureau of Economic Analysis (DoC)

BJS- Bureau of Justice Statistics (DoJ)

BTS- Bureau of Transportation Stat. (DoT)

Census- DoC

EIA- Energy Information Admin. (DoE)

ERS- Economic Research Service (DoA)

NASS- Nat. Agricultural Stat. Service (DoA)

NCES- Nat. Center of Education Stat. (DoE)

NCHS- Nat. Center for Health Stat . (HHS)

NCSES- Nat. Center for Science and Engineering Sat. (NSF)

ORES- Off. of Research, Evaluation, and Statistcs (SSA)

SOI- Statistics of Income (IRS)

Image from AmStat (permission pending)

29 of 37

One Statistical Office in US: Why Not?

1. Privacy: The Privacy Act of 1974, Confidential Information Protection and Statistical Efficiency Act of 2002 (CIPSEA), and Statistical Policy Directive No. 1 (2014) require agencies to ensure that the collection and maintenance of citizens' data is accurate, confidential, and within legal restrictions. With different offices having access to those records, there would be less possibility of everything being leaked.

2. Security: Along the lines of fewer offices having access to data records. The more servers that hold the data, the safer it is. The times when an exchange of information is necessary laws and regulations among departments allow to protect access to data.

3. Integrity: The income you report to IRS might be different from what you report to the Census Bureau.

4. Methodology: Sometimes data must have a higher number of people questioned so the accuracy will be better; different methods of collection or sampling may be required.

5. Popularity: Anything being done by the government has a political dimension, especially funding for employees and for modernizing and updating technology, attractiveness of the research, and repetition of statistical programs by agencies.