Wikipedia Pageviews Monitor: Top Articles & Spikes avatar

Wikipedia Pageviews Monitor: Top Articles & Spikes

Pricing

from $0.21 / 1,000 view records

Go to Apify Store
Wikipedia Pageviews Monitor: Top Articles & Spikes

Wikipedia Pageviews Monitor: Top Articles & Spikes

Track Wikipedia article views by day or month. Get daily top articles, rolling spike ratios and per-article totals, averages and peaks. No API key; up to 50 articles and 90 days.

Pricing

from $0.21 / 1,000 view records

Rating

0.0

(0)

Developer

Cliqto Media

Cliqto Media

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

2 days ago

Last modified

Features

Limited permissions

Share

Get daily or monthly Wikipedia article views, daily top articles and interest spikes from Wikimedia metrics. Add exact article titles; no API key is needed. The three-article, 30-day sample costs $0.057 at the base tier. Unavailable source data is marked clearly; a spending limit can produce a partial result.

Wikipedia Pageviews and Trends Monitor: article input, view chart, totals and peaks

Table of contents

What it does

  • Articles: one row per article and UTC day, or per calendar month. Get total views, average, peak date and spike counts in the run summary.
  • Top: get the most-viewed pages for one Wikipedia project and UTC day. Keep the source ranks, including ties and rank gaps.
  • Spikes: compare each point with the average of the previous points. The default baseline is seven points; a spike starts at twice that average.

Views measure reading activity. They do not measure unique readers, sales or buying intent. Top lists include main pages and namespaces such as Special:. This Actor does not remove them or change their rank. It does not combine redirects, old article names or visits across titles.

Quick start

  1. Open Input.
  2. Keep the three sample articles: Artificial intelligence, OpenAI and Python (programming language), on en.wikipedia.
  3. Keep mode as articles and granularity as daily. Leave dates blank for the recent 30-day window.
  4. Click Start.
  5. Open Pageviews. Check article, date and views. Open Run summary for totals and peaks.

Expected result: 90 view points when the source has all 30 days for each article. The window ends two UTC days ago to allow time for source updates. The sample is a UI prefill; an empty API input {} is invalid.

Input

Input fields

FieldTypeRequiredRuntime default / UI prefillValues and effectRecommendation
modestringNoarticlesarticles or topUse articles for history; top for a daily ranking.
articlesarray of objectsIn articles modeNo runtime default; UI prefill has three articles1–50 objects, each with title and optional project. Title is an exact Wikipedia title, at most 255 characters. Spaces and underscores are equivalent. Identical project/title pairs are removed.Copy the page title; do not paste a URL.
projectstringNoen.wikipediaWikipedia language project for top and articles without their own project. en.wikipedia.org is also accepted.Set project in each article for mixed languages.
startDatestringNo29 days before endDateInclusive YYYY-MM-DD, from 2015-07-01For monthly, use the first day of a month.
endDatestringNoTwo UTC days agoInclusive YYYY-MM-DD, no later than yesterday; at most 90 inclusive days after startDateFor monthly, use the last day of a month.
datestringNoTwo UTC days agoYYYY-MM-DD, top mode only, from 2015-07-01 to yesterdayA missing source day is reported; it is not silently changed.
accessstringNoall-accessall-access, desktop, mobile-web, mobile-appKeep all-access for combined access channels.
agentstringNouserArticle mode: user, all-agents, spider, automatedKeep user for human-classified traffic. Top has no agent filter; keep the default.
granularitystringNodailydaily or monthly; top requires dailyMonthly requires complete calendar months within 90 days.
spikeThresholdnumberNo21.1–100; views divided by the rolling average must reach this multiplierIncrease it to flag larger changes.
spikeWindowintegerNo71–30 previous points; days for daily, months for monthlySet 1 for a month-to-month comparison.
maxResultsintegerNo45001–4500 saved rows, including not-found checksLower it to limit output. Top returns at most 1000 source entries.

Empty strings, null field values, empty articles in articles mode, unknown keys and invalid dates are rejected. Article titles may include Unicode. No content or contact details are collected.

Interacting settings

SettingsBehavior
Article project + global projectArticle project wins; the global project supplies missing article projects.
endDate without startDateStart is 29 days before the supplied end date.
monthly + datesOnly full calendar months are allowed. Counts are for the full month; date is its first day.
spikeWindow + date windowOnly returned earlier points are used. No extra baseline dates are fetched. The first spikeWindow points have null spikes.
top + series settingsArticles and start/end dates do not select top entries. Keep agent=user and granularity=daily. All supplied dates and objects must still be valid. Spikes and article totals are not calculated for top.
maxResults + maximum chargeThe first reached limit stops saving rows. The summary names the reason.

Input examples

Three daily points with a one-point baseline:

{
"mode": "articles",
"articles": [
{ "project": "en.wikipedia", "title": "Artificial intelligence" }
],
"startDate": "2026-09-01",
"endDate": "2026-09-03",
"spikeWindow": 1
}

This returns three daily points and an article total of 39,592 views. Base cost: $0.0309.

Daily top list:

{
"mode": "top",
"project": "en.wikipedia",
"date": "2026-09-01",
"maxResults": 1000
}

That source list has 999 entries. Main Page is rank 1 with 6,785,879 views. Base cost: $0.3297. The source may return fewer than 1000 entries.

Monthly points:

{
"articles": [{ "title": "Artificial intelligence" }],
"startDate": "2026-08-01",
"endDate": "2026-09-30",
"granularity": "monthly",
"spikeWindow": 1
}

Two complete months return 355,079 and 540,333 views. Base cost: $0.0306.

Output

Output example

The first row for the daily input above is shown below. The fetch time is from a real run. The first point has no previous baseline, so its spike values are null.

{
"project": "en.wikipedia",
"article": "Artificial intelligence",
"date": "2026-09-01",
"views": 12087,
"access": "all-access",
"agent": "user",
"granularity": "daily",
"rank": null,
"status": "complete",
"reason": null,
"spike": null,
"spikeRatio": null,
"baselineAverage": null,
"provider": "Wikimedia Analytics API",
"licence": "CC0-1.0",
"sourceUrl": "https://wikimedia.org/api/rest_v1/metrics/pageviews/per-article/en.wikipedia/all-access/user/Artificial_intelligence/daily/2026090100/2026090300",
"articleUrl": "https://en.wikipedia.org/wiki/Artificial_intelligence",
"fetchedAt": "2026-10-14T00:38:24.694Z",
"sourceUpdatedAt": null
}

Dataset fields

FieldTypeNull / emptyMeaning
projectstringNoNormalized project, such as en.wikipedia.
articlestringEmpty on an unavailable top listTitle with spaces.
datestringNoUTC day; first day of the month for monthly counts. On not_found, the requested start date or top date.
viewsinteger or nullNull on not_foundSource view count. Zero is kept as zero.
accessstringNoRequested access channel.
agentstring or nullNull for topRequested article traffic agent. Top endpoint has no agent filter.
granularitystringNodaily or monthly.
rankinteger or nullNull for articles or not_foundSource top rank; ties and gaps are kept.
statusstringNocomplete or not_found.
reasonstring or nullNull for completeNO_DATA_FOR_TITLE_PROJECT_OR_DATE on a source 404.
spikeboolean or nullNull for top, unavailable or missing/zero baselineTrue when spikeRatio reaches spikeThreshold.
spikeRationumber or nullNull without a positive full baselineViews divided by the mean of the previous spikeWindow points.
baselineAveragenumber or nullNull without enough previous points or for topRolling mean; a zero baseline stays zero.
providerstringNoWikimedia Analytics API.
licencestringNoCC0-1.0 for the numeric metrics.
sourceUrlstringNoMetrics request URL for this query.
articleUrlstringEmpty on an unavailable top listLink to the exact title, without resolving redirects.
fetchedAtstringNoUTC time of reading the result.
sourceUpdatedAtstring or nullAlways nullSource endpoint does not provide a revision timestamp.

Tables and run records

  • Pageviews table: project, article, date, views, access, agent, granularity, rank, status and articleUrl.
  • Spikes table: project, article, date, views, baselineAverage, spikeRatio, spike and status. It shows all points, including those without a spike.
  • RUN_SUMMARY: status, reason, provider, licence, mode, rows, usefulRows, notFound, requests, retries, bytes, startedAt, finishedAt and articleSummaries.
  • OUTPUT: status, rows, reason and articleSummaries.

Each articleSummaries item has project and article (strings), points, total, peakViews and spikes (integers), average (number), peakDate (UTC date string), complete (boolean) and missingDates (array of UTC date strings). Totals use saved points only. An item is absent for unavailable articles and in top mode. A truncated article, or one with dates the source does not have, has complete=false; totals are for that delivered subset.

RUN_SUMMARY.status is complete, partial, not_found or error. reason is null or an error/limit code. Row and request counters are integers; timestamps are UTC strings. A not_found result has one paid check row per unavailable query, with null views. It does not prove that the page does not exist.

Pricing and limits

One automatic start costs $0.03. Each saved daily/monthly point, top entry or not-found check is one view-record event. Article summaries have no extra charge.

TierPer 1000 saved rowsPer row
FREE / base$0.30$0.000300
BRONZE$0.285$0.000285
SILVER$0.270$0.000270
GOLD / PLATINUM / DIAMOND$0.255$0.000255

The FREE tier is the base price, not a promise of free rows. Standard run platform usage is included in the event price.

ScenarioSaved rowsBase totalGOLD total
Unavailable article check1$0.0303$0.030255
Small series10$0.033$0.03255
Three-article sample, 30 days90$0.057$0.05295
Series or top subset100$0.060$0.05550
Top list with 1000 entries1000$0.330$0.285
50 articles × 90 days4500$1.380$1.1775
Runtime input error before rows0$0.030$0.030

Set Maximum cost per run to control events. The minimum is $0.03. At that limit, the start can use the budget and no source requests are made. A higher limit may return fewer points. Saved rows and the start remain charged on failed, partial, repeated or cancelled runs. An input rejected before a run starts has no Actor start event.

Use cases

JobInputUseful output
Follow reading activity for a brand or topicExact titles in articlesdate, views, total, average, peakDate
Prepare a daily editorial topic listtop + project + datearticle, rank, views, articleUrl
Compare language editionsPer-article projectsproject, article, date, views
Flag changes in reading activityspikeWindow + spikeThresholdbaselineAverage, spikeRatio, spike

Repeated runs and scheduling

Every run reads the requested metrics again and charges for saved rows. No baseline, only-new filter or rank movement history is kept between runs. Wikimedia can revise historical data.

To monitor recent activity, save your input in an Apify task and set a schedule yourself. Use the Spikes table or store each run's output in your own system. This Actor does not create schedules or webhooks.

API, export and integrations

Start it through the Apify API with your own token in the Authorization header:

curl -X POST 'https://api.apify.com/v2/actors/cliqtomedia~wikipedia-pageviews-monitor/runs?memory=256&timeout=300&maxTotalChargeUsd=0.10' \
-H 'Authorization: Bearer YOUR_APIFY_TOKEN' \
-H 'Content-Type: application/json' \
-d '{"articles":[{"title":"Artificial intelligence"}],"startDate":"2026-09-01","endDate":"2026-09-03"}'

Poll the returned run ID, then read its Dataset and RUN_SUMMARY. Export the flat Dataset as JSON or CSV. Use project, article and date to join rows in your spreadsheet or data pipeline.

Use with AI agents (MCP)

Open this Actor in Apify MCP. Apify authentication is required to run it.

Small UI sample input:

{
"articles": [
{ "project": "en.wikipedia", "title": "Artificial intelligence" },
{ "project": "en.wikipedia", "title": "OpenAI" },
{ "project": "en.wikipedia", "title": "Python (programming language)" }
]
}

Check project and article for identity, date for the UTC period, views for the count and spike for the rolling alert. The 90-point sample costs $0.057 at base tier and normally takes under 30 seconds. Respect the limits of 50 articles, 90 inclusive days and 4500 rows; larger work may need polling. The Actor uses limited permissions and pay-per-event pricing.

Limits and partial results

  • Wikipedia language projects only. No country charts, article text, unique readers or edit counts.
  • Up to 50 input articles and 90 inclusive days. Monthly mode needs full months; partial months are rejected.
  • Top lists are for one project/day, with up to 1000 source entries. They include namespaces and may contain fewer entries.
  • user is the default series agent. Top has no configurable agent filter.
  • Source data starts in July 2015 and arrives after the UTC day ends. Two-day lag is a safe default, not a source availability guarantee.
  • Renames and redirects remain separate titles. Counts can differ from a tool that combines redirect views.
  • Insufficient or zero spike baseline returns null, not infinity or an invented spike.
  • maxResults or charge limit: run succeeds with partial and MAX_RESULTS or CHARGE_LIMIT.
  • Dates the source does not have (before an article was created or renamed, or not yet processed): no row is written for them, and they are not filled with zero. They are listed in missingDates; the run succeeds with partial and SOURCE_DATES_MISSING. Spike baselines need the preceding periods to be present.
  • Source, format, deadline or graceful cancellation: run fails. Saved rows remain; summary is partial if rows exist, otherwise error.
  • Internal source work is bounded to 240 seconds, 160 attempts, 16 MiB per response and 64 MiB per run. A shorter platform timeout reduces work. Use 256 MB memory.
  • A forced platform kill can stop the process before it writes final diagnostics. Dataset writes and event charges are separate platform operations; no exactly-once guarantee is made after a lost storage response.

Troubleshooting

SymptomCheckAction
INVALID_INPUTExact titles, object arrays, date format/order, supported project and numeric limitsCorrect input. Monthly needs full calendar months. Empty API input is invalid.
not_found with null viewsTitle, project and requested dateTry a known earlier date or the exact current/old title. Do not read null as zero.
partial / MAX_RESULTSSaved row count and article summary complete flagRaise maxResults within the 4500 limit or split the input.
partial / CHARGE_LIMITMaximum cost and saved row countRaise the spending cap for a rerun; the rerun bills again.
partial / SOURCE_DATES_MISSINGmissingDates in the article summaryNew or renamed article, or a very recent date. Gaps are not zero views.
SOURCE_ERROR or DEADLINERun summary and smaller known inputRetry later or reduce input size. Do not treat failure as no interest.
SOURCE_FORMATSource response has an unexpected shape or identityReport the run ID.
No spikesBaseline length, zero baseline and multiplierUse enough periods or reduce spikeWindow; top does not calculate spikes.

FAQ

Do I need a Wikimedia key or proxy? No.

Does zero mean the same as null? No. Zero is a supplied count. Null means no available metric or no valid baseline.

Does top mean fastest-growing? No. It means most-viewed for the selected day. Use article series for spike ratios.

Can I ask for yesterday? Yes, but data may not yet be available. Missing days are listed in missingDates and the run ends as partial. The default ends two days ago.

Are redirects combined? No. Query each exact title you want to measure.

Does a spike predict sales or searches? No. It reports a change in Wikipedia view counts only.

Does repeating the same input cost money? Yes. Each run starts again and charges for saved results.

Source and support

Source: Wikimedia Analytics API. Numeric data is available under CC0 1.0, as stated in the access policy. The Actor reads and reshapes these counts; Wikimedia does not endorse this product.

For a bug or question, use the Actor's Issues tab on Apify. Include the run ID, mode, title/project, requested dates and a short error code. Send no tokens, cookies or billing details.