TREND RESEARCH From published sources / US · UK · Canada · Australia
RESEARCH METHODS / TRANSPARENCY
A measurement is not a market. Here is how we tell the difference.
SurgeAtlas documents what a source observed, which population it covers and when the reading was collected. Models help frame questions, but they cannot manufacture customers or future outcomes.
Wikipedia tells us about article reading. GitHub shows work on a repository. Google Trends describes relative search interest. These are different signals—not competing estimates of the same audience.
Every source below reflects a specific behavior and has explicit limits. Availability may differ by topic, region or collection date.
Original data source
Search interest
Relative interest index
Query attention scaled within the requested region and period. It is not monthly search volume, a purchase count or evidence of market share.
WHAT TO CHECK
Confirm the exact keyword, time frame and geography before comparing.
Original data source
Wikipedia readership
Human pageviews
Daily readings of a specific article from the official Wikimedia pageview endpoint. Global article readership cannot establish country-specific purchasing intent.
WHAT TO CHECK
Check the article title, collection date and any gaps.
Original data source
Open-source activity
Repository activity
Stars and repository events can illuminate developer curiosity, not commercial installations or paid seats. Release announcements may concentrate activity.
WHAT TO CHECK
Check project maintenance, issue trends and whether the observed repository matches the concept.
Original data source
Community discussion
Sampled conversations
A bounded set of posts or comments, not a complete census of users or buyers. News coverage may drive multiple simultaneous discussions.
WHAT TO CHECK
Look at community selection, sample window and evidence of actual workflows.
FROM DATA TO A PAGE
What happens between a source and a public page?
These are safeguards, not guarantees of perfect accuracy. Missing or stale data may delay publication.
Collect readings with a known source.
Save where each reading came from, when it was recorded and what it counted. A pending task is not yet real data.
Check that the history is useful.
Check that there is enough history to make a fair comparison. Hold back incomplete charts.
Keep estimates separate from facts.
Show past readings as facts. Clearly label any future outlook as an estimate.
Publish only what the evidence supports.
Show real observations when available. More detailed articles are published only after additional research and checking.
INTERPRETATION RULES
Four boundaries that protect a decision.
Why do some topics show no score?
A missing score means insufficient qualifying data or source access. It does not indicate a score of zero or that a market has no demand.
Is a six-month forecast a prediction of sales?
No. It is a statistical projection of a source-specific signal. There is no universal conversion from attention to revenue.
Are Google Trends and Wikipedia results comparable?
They can support different research questions. A normalized Google interest index and a global Wikipedia pageview total cannot be added or interpreted as a shared audience size.
Does an AI-written analysis count as verified fact?
No. Structured validation checks completeness and evidence formatting. Editorial interpretation and provider-generated statements still require independent scrutiny.
GO DEEPER
Good methods are useful only when you apply them.
Use the research playbook to inspect a chart, challenge a source and test a specific hypothesis in the real world.