Confounds and consequences in geotagged twitter data
Open accessAbstract
Twitter is often used in quantitative studies that identify geographically-preferred topics, writing styles, and entities.These studies rely on either GPS coordinates attached to individual messages, or on the user-supplied location field in each profile.In this paper, we compare these data acquisition techniques and quantify the biases that they introduce; we also measure their effects on linguistic analysis and textbased geolocation.GPS-tagging and selfreported locations yield measurably different corpora, and these linguistic differences are partially attributable to differences in dataset composition by age and gender.Using a latent variable model to induce age and gender, we show how these demographic variables interact with geography to affect language use.We also show that the accuracy of text-based geolocation varies with population demographics, giving the best results for men above the age of 40.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In location-based data collections, different strategies to match locations, such as using message geo-tags or the author's self-declared location, affect the user demographics and contents of a sample.
“different strategies to match locations, such as using message geo-tags or the author's self-declared location, affect the user demographics and contents of a sample (Pavalanathan and Eisenstein, 2015)”
From Social Data: Biases, Methodological Pitfalls, and Ethical Boundaries (2019) · cites this paper as Pavalanathan and Eisenstein, 2015
The accuracy of text-based geo-location of Twitter users varies across samples, based on whether either the user-supplied location or their tweets' GPS coordinates were used as proxies for user location.
“the accuracy of text-based geo-location of Twitter users varies across samples, based on whether either the user-supplied location or their tweets' GPS coordinates were used as proxies for user location (Jurgens et al., 2015b; Pavalanathan and Eisenstein, 2015).”
From Social Data: Biases, Methodological Pitfalls, and Ethical Boundaries (2019) · cites this paper as Pavalanathan and Eisenstein, 2015
The performance of existing inference models is sensitive to various user-related confounds such as age or gender.
“several empirical studies show that the performance of existing inference models is sensitive to various user-related confounds such as age or gender (Pavalanathan and Eisenstein, 2015;Landeiro and Culotta, 2016).”
From Social Data: Biases, Methodological Pitfalls, and Ethical Boundaries (2019) · cites this paper as Pavalanathan and Eisenstein, 2015
Even a very accurate model may introduce systematic errors concentrated on certain classes of messages or of users.
“Even a very accurate model may introduce systematic errors concentrated on certain classes of messages or of users (Cohen and Ruths, 2013; Pavalanathan and Eisenstein, 2015; Tramer et al., 2015).”
From Social Data: Biases, Methodological Pitfalls, and Ethical Boundaries (2019) · cites this paper as Pavalanathan and Eisenstein, 2015
Cited by
Sorted by citations
Fake news on Twitter during the 2016 U.S. presidential election
Grinberg, N., Joseph, K., Friedland, L., Swire-Thompson, B., & Lazer, D. (2019) · Science · doi:10.1126/science.aau2706 · Cited by 1,803 · see citing works
Social Data: Biases, Methodological Pitfalls, and Ethical Boundaries
Olteanu, A., Castillo, C., Diaz, F., & Kıcıman, E. (2019) · Frontiers in Big Data · doi:10.3389/fdata.2019.00013 · Cited by 763 · see citing works
“(see, e.g., Cohen and Ruths, 2013; Nguyen et al., 2014; Jurgens et al., 2015b; Pavalanathan and Eisenstein, 2015).”
BackgroundMethodInfluentialStudying user income through language, behaviour and affect in social media
Preoţiuc-Pietro, D., Volkova, S., Lampos, V., Bachrach, Y., & Aletras, N. (2015) · PLOS ONE · doi:10.1371/journal.pone.0138717 · Cited by 219 · see citing works
Social Data: Biases, Methodological Pitfalls, and Ethical Boundaries
Olteanu, A., Castillo, C., Diaz, F., & Kiciman, E. (2016) · doi:10.2139/ssrn.2886526 · Cited by 763 · see citing works
- Authors
- Pavalanathan, U., & Eisenstein, J.
- Affiliations
- Georgia Institute of Technology (US)
- Venue
- Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing
- Published
- 2015
- Pages
- 2138-2148
- OpenAlex
- W1754225409
- Cited by
- 104
- References
- 0
- Citations by year
- 2026: 5 · 2025: 5 · 2024: 6 · 2023: 6 · 2022: 6 · 2021: 14 · 2020: 9 · 2019: 10 · 2018: 13 · 2017: 17 · 2016: 11 · 2015: 2
- FWCI
- 8.08
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Conference on Empirical Methods in Natural Language Processing (EMNLP) (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Science Foundation (1111142, 1452443)
- Georgia Institute of Technology
- National Institutes of Health
- National Institute of General Medical Sciences (R01GM112697)
- Air Force Office of Scientific Research
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Authorship Attribution and Profiling
- Social Sciences › Transportation › Human Mobility and Location-Based Analysis
- Social Sciences › General Social Sciences › Computational and Text Analysis Methods
- Keywords
- geolocation accuracy, data bias, age and gender differences, latent variable models, linguistic analysis
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
