1 of 10

How to Find &

Evaluate Data

Stat 1601

Fall 2023

2 of 10

Hello!

Laura Hjerpe

Research Data Management Librarian

UVA Library

lhjerpe@virginia.edu

Erich Purpur

Engineering Librarian

UVA Library

epurpur@virginia.edu

2

3 of 10

1. Evaluating Data

3

4 of 10

Different kinds of data

Designed data

Surveys
Experiments

Administrative Data

“Business practice data”
IRS data for individuals and businesses
Patent databases

Opportunity data

GPS and sensors
Social media

Procedural data

Compensation policies
Affordable Care Act

4

I really like the types of data defined by Keller, Lancaster, & Shipp (2017):

"Designed Data are generated in the pursuit of scientific discovery. Designed data include statistically designed data collections, such as data generated from: surveys, experimental designs, registries, and intentional observational collections.

Administrative Data, also referred to as “business practice” data, are collected for the administration of an organization, program, or service processes. These data provide an opportunity for gathering information that exists due to normal economic and social activity. Examples of administrative data include Internal Revenue Service data for individuals and businesses, Social Security earnings records, patent and trademark databases, Medicare and Medicaid health utilization data, banking and other financial data, industrial production processes, such as tracking supply chains end-to-end (Pires et al. 2017), taxi trip data, and local data generated from 911 calls and Emergency Management Services (EMS) responses, property assessment and tax data, and data from health and human services, parks and recreation, libraries, and environmental services (e.g., trash and recycling, water and utilities, projects and planning, transportation, and building permits).

Opportunity Data are data generated on an ongoing basis as society moves through its daily paces. Opportunity data are derived from a variety of sources such as GPS systems and embedded sensors, social media exchanges, mobile and wearable devices, and Internet entries. Captured through a variety of methods including direct flows, Internet searches, web crawling and scraping, these data may exist in a variety of electronic and physical modalities.

Procedural Data are data derived from policies, procedures, and legal requirements; they are the rules and regulations that govern and shape our lives. These policies and procedures affect our work, our personal lives, and society. Examples of procedural data include compensation policies, the Affordable Care Act, the Department of Defense policy “Don't Ask, Don't Tell,” and Supreme Court rulings."

##Launch Poll3

5 of 10

Questions to Ask Yourself

Who collected the data? Is this a credible source?
Did the answer come from a valid process and methodology?
What kind of description of the data is there?
Are the data readily accessible?

5

*** Remember to show kaggle and why it is potentially a bad data source

Overall Data Literacy:

Concept: Authority is Constructed and Contextual.

Which organization is most likely to collect data? If you can figure out who might collect data, that can help you figure out where to start your search. Who would care enough to collect the data? Who would reasonably have access to the data or the resources to collect the data?

Concept: Information Creation as a Process.

What process was used to create the data? Does the process of data collection introduce concerns (i.e., web scraping which leads to concerns about authoritative nature of data – on the other hand maybe it is possible to replicate the data gathering?)? Ask yourself, Did the answer come from a valid process and methodology? Is there a better process (accessible to you)?

Jenn will use Kaggle: Florence Nightingale’s Data on Deaths: https://www.kaggle.com/datasets/ukveteran/florence-nightingales-data-on-deaths

ICPSR: The Common Sense Census: Media Use by Tweens and Teens in America, A Common Sense Media Research Study, United States, 2015: https://www.icpsr.umich.edu/web/ICPSR/studies/38018

Compare ICPSR and Kaggle [10 minutes]

Compare ICPSR and Kaggle. Ask yourself the following questions.

Where do the data come from? Is this a credible source?
What kind of metadata do you get? Description of overall dataset, about variables, how data were collected, over what period, etc.
Are the data accessible?

Questions

What did you notice when comparing ICPSR and Kaggle?
How to think through results that are too broad and/or too narrow

Too narrow: Just because you can imagine it, doesn’t mean it exists (or maybe not in the format you want it to). And this seems to be the more common problem.
Too broad: Is there anything more specific, something you are really interested in. Is there an aspect on your topic that you want to focus on? If there is a lot of data on your area of interest, does a subset of this area have data available?

6 of 10

Other things to consider

Access

Format

Documentation

6

-Look for Star Trek Scripts in Kaggle

Let’s say you have made progress on finding a possible dataset. What are the next steps? What are the kinds of things I look for when I’m reviewing a new dataset?

Access – are the data open and readily available? Open but requiring a data use agreement? Or is it proprietary, likely requiring paying a fee, plus signing a data use agreement and/or a license?

Proprietary data will not be a good choice when facing short, semester-based deadlines.

Format – what format is the dataset in? Is it already in a format that will enable easy import into statistical software?
Documentation – what kind of methods documents are available? This is especially important when looking at a secondary datasets, especially surveys. The documentation should answer questions like:

how did the researchers collect the data?
What do the variables represent?
How do the researchers treat categorical variables, or missing data?
What are the units of measurement
ALWAYS read the documentation.

Importantly, for this assignment, you need to match two datasets to one variable. What variable will that be? Geography (county, state, country) is a good choice, since many datasets will include geography. You need to make sure you can match on something specific.

7 of 10

2. Useful Data Sources to Consider

7

8 of 10

There is a research guide!

https://guides.lib.virginia.edu/stat1601

8

[25 minutes]

There is a Research Guide: https://guides.lib.virginia.edu/stat3220. Let’s follow along there.

Library licensed:

ICPSR is the social science data repository that has very well curated data, and by curated I mean that everything is well labeled and there is sufficient documentation for every dataset. The search functionality is excellent, making it easy to find what you are looking for. You (usually) need to create an account to download (UVA is a member institution).
Social Explorer is the best platform for finding US Census data quickly and easily. The interface is so much nicer than census.gov and historical data is available as well. There is other data available as well, such as data on religion and crime
Data-Planet is another great place to find data. This is mostly from open/administrative/governmental sources, but we also have some specially licensed market data there as well. This is the one that I mentioned that downloading the data would be tricky here; I need to ask the vendor about the API.
Here is the complete list of the data licensed by the library. The highlights are the historical business data, election data, and Gallup survey microdata.

Government data

Data.gov or try Data & Statistics Research Guide: https://guides.lib.virginia.edu/data
State and local open data portals. https://opendata.charlottesville.org/ and https://data.virginia.gov/
International: Our World in Data https://ourworldindata.org/

Some news sources (good for relevance):

ProPublica: https://www.propublica.org/datastore/
Pew Research: https://www.pewresearch.org/tools-and-resources/
News sources’ GitHub: NYTimes, Washington Post, BuzzFeed News

Academic literature.

Every empirical academic paper is going to have a methods section. Look for articles similar to a topic you are interested in, and check the methods section.
Example: Injury Rates in Major League Baseball During the 2020 COVID-19 Season. The very first line of the Methods section states, “Data from the 2018-2020 MLB transaction reports were extracted online at mlb.com/transactions.”

Start in one place and keep iterating. Use the first place to see if sources are cited, and work from there.

Statista example: Covid Vaccine….search for sources for USA…OWID and CDC, look at References, see Our World in Data… OWID points to CDC page with data.
PolicyMap, too.

9 of 10

Thanks!

Any questions?�You can find me at:

lhjerpe@virginia.edu
epurpur@virginia.edu

9

10 of 10

Slide Template Credits

Special thanks to all the people who made and released these awesome resources for free:

Presentation template by SlidesCarnival
Photographs by Unsplash
Template

10