Creating full individual-level location timelines from sparse social media data
Bibliographic record
Abstract
In many domain applications, a continuous timeline of human locations is critical; for example for understanding possible locations wherea disease may spread, or the flow of traffic. While data sources such as GPS trackers or Call Data Records are temporally-rich, they are expensive, often not publicly available or garnered only in select locations, restricting their wide use. Conversely, geo-located social media data are publicly and freely available, but present challenges especially for full timeline inference due to their sparse nature. We propose a stochastic framework, Intermediate Location Computing (ILC) which uses prior knowledge about human mobility patterns to predict every missing location from an individual's social media timeline. We compare ILC with a state-of-the-art RNN baseline as well as methods that are optimized for next-location prediction only. For three major cities, ILC predicts the top 1 location for all missing locations in a timeline, at 1 and 2-hour resolution, with up to 77.2% accuracy (up to 6% better accuracy than all compared methods). Specifically, ILC also outperforms the RNN in settings of low data; both cases of very small number of users (under 50), as well as settings with more users, but with sparser timelines. In general, the RNN model needs a higher number of users to achieve the same performance as ILC. Overall, this work illustrates the tradeoff between prior knowledge of heuristics and more data, for an important societal problem of filling in entire timelines using freely available, but sparse social media data.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.008 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".