Frequently Asked Questions about Information Tracer


Q: What is Informaiton Tracer?
A: Information Tracer is a tool to collect, contextualize, and investigate how information spreads across social media platforms.

Q: What can I do with Informaiton Tracer?
A: People have used Informatin Tracer to:

Q: What can I search?
A: Keyword, URL, hashtag, or a combinaiton of those.

Q: How to search?
A: You can use three boolean operators -- AND OR NOT (must be all-cap). For example, ("Ukraine" AND "NATO") OR ("Ukraine" AND "European Union") means any post with "Ukraine" and "NATO", or with "Ukraine" and "European Union". Use double quotations to exact-match multile words. Also logics inside the same parenthesis have precedence over logics outside.

Q: Can I search in any language, and during any time?
A: Yes.

Q: What platforms do Information Tracer cover?
A: We currently collects data from Twitter, Youtube, Reddit, Facebook, Instagram.

Q: How much data does Information Tracer collect?
A: For each query, IT collects up to 1000 Facebook posts, 1000 Instagram posts, 1000 Reddit posts, 1000 Twitter posts, and 1000 YouTube posts. IT collects posts with the highest number of interactions (engagements) first.

Q: Why those limitations?
A: We do not want to burden the platform and aim to be efficient. A small number of posts often cover most of the interactions. It is not informative to collect posts with 0 interactions, because they have no exposure. Therefore, our aggregated statistics closely estimates the true impact of a narrative (keyword, URL, etc,.).

Q: What if I want to collect all posts during a time period?
A: You can reduce the search window, query multiple times, and combine results.

Q: What is the “information quality score” you generate?
A: The lower the quality score, the more “coordianted” the spread patter is. By default, the score is 100. We calculate multiple metrics to quantify how “coordinated” or “repetitive” the information spread pattern is. You can find those metrics under the tab “Suspicious patterns.” For each metric, we set a threshold, above which the metric is considered abnormal. For each abnormal metric, we deduct 10 points from the quality score.

Q: My YouTube results do not match my query. Why is it?
A: Since of May 12, 2023, we have resolved this problem. Now all Youtube results (based on titile, channel name, or description) match the query. Old answer: The YouTube API returns results based on a probability model. As a result, not all YouTube results match your query. Some results are “probably related”, but might be irrelevant. Use YouTube results with extra caution.

Q: What about other platforms – TikTok, Whatsapp, Telegram?
A: We are working on adding those platforms. Stay tuned.

Q: What about private posts and private accounts?
A: We do not have access to private posts or private accounts.

Q: I want to try Information Tracer. What should I do?
A: To get started, please send an email to zhouhan@safelink.network. Please explain your use case. If the use case is legitimate, we will grant you a 30-day trial for Information Tracer.

Q: Where can I learn more about Information Tracer? Are there any presentation slides?
A: You can find recent talk we presented to 2023 World Press Institute Fellows here.