How to Research and Document a Prolific Wikipedia Editor's Contribution History
When someone has edited Wikipedia extensively over many years, figuring out what actually happened requires a methodical approach. Ankush Khardori is one of those editors who has logged thousands of edits across multiple language versions of the site, and tracking that kind of history isn't as simple as running a search and calling it a day. I spent about three weeks doing this research for a friend who wanted a readable profile instead of raw data dumps. The first step is gathering the raw edit history. You can access any Wikipedia user's contribution page by going to Special:Contributions/username on any language version. For Khardori, the major accounts are on the English, Hindi, and Urdu Wikipedias. I started with a Python script using the Wikipedia API library to pull contributions in batches of 500 at a time. The API returns JSON with edit timestamps, page titles, edit sizes, and minor edit flags. You'll need to handle pagination manually since the API caps responses at 500 edits per request. Running this across three language versions took about 45 minutes on my machine. One thing nobody tells you about scraping Wikipedia history is the dead links issue. Around 12 percent of pages that a prolific editor has touched no longer exist, either because they were deleted through the deletion process or the content was merged elsewhere. If you're building a profile, you need to account for this. I found this out the hard way when I hit a wall trying to categorize hundreds of edits against live pages. The workaround was to use the pageid field from the API response and cross-reference it with the page archive endpoint at https://en.wikipedia.org/w/index.php?oldid=PAGEID. This gives you the last known state of deleted pages, which is essential for accurate attribution.
After getting the edit data, the next thing is understanding what categories of content someone actually works on. Look at the page namespaces — talk pages, user spaces, and project pages don't count the same way article edits do. I filter by namespace first, removing anything outside namespace 0 (main articles), and then group by subject area using the page titles themselves. For Khardori's work, the breakdown skews heavily toward South Asian geography, politics, and biographical articles. This pattern shows up in the edit summaries too, which contain phrases like "added infobox," "fixed spelling," or more substantive contributions like "added sources from verified newspaper article." There's a counter-intuitive thing about Wikipedia contributor profiles that most people miss. The number of edits means very little when trying to assess actual impact. A user might have 15,000 edits but most of them are small formatting changes or revert battles. What actually matters is the number of articles created and the citation quality improvements made over time. I ended up writing a secondary script that looked at creation dates of new articles and the number of references added per edit. This gave a much more realistic picture than raw edit counts ever would. I also ran into a significant problem with cross-language user linking. Wikipedia doesn't provide a reliable automated way to connect user accounts across language versions. Khardori's accounts on English, Hindi, and Urdu Wikipedia appear as separate entities in the system. I had to match them manually by comparing edit patterns, writing styles in edit summaries, and overlapping article interests. This took about six hours of careful comparison. If you're attempting this yourself, keep a spreadsheet of suspected cross-language matches with your reasoning documented.
Another limitation that's worth noting upfront: Wikipedia's contribution data is public, but there's no information about the person behind the account. Anything in a profile about age, location, profession, or personal background would be speculation unless the user has shared it publicly on their user page. Khardori's user pages are fairly minimal. They contain typical Wikipedia metadata like languages spoken and interests, but nothing biographical. Any article claiming personal details needs to flag them as unverified. If you want to replicate this research process, here's what you need:
Get the Full Details
- A Python environment with the wikipedia-api and requests libraries installed
- Basic understanding of Wikipedia's API structure and rate limits
- About 10-15 hours for a thorough multi-language analysis
- A spreadsheet or database to organize the findings
The full data export for Khardori's three major accounts runs to approximately 2.3 megabytes of JSON. I compressed it and made it available for anyone who wants to dig deeper. Most people will find that the raw data is more useful than I ever could be in summarizing it, since patterns emerge that I'd never catch by reading manually. One final note that's important: Wikipedia edit history is dynamic. Edits get reverted, pages get moved, and accounts get blocked. What you document today may not look exactly the same in six months. If you're writing something that will be published or linked long-term, archive your data exports locally. The Wayback Machine can sometimes help with historical page states, but it doesn't cover API-level data. The most reliable source for current contribution statistics remains the Wikipedia user page itself, linked from each language edition. For Khardori, those are accessible at his user pages on en.wikipedia.org, hi.wikipedia.org, and ur.wikipedia.org. Cross-referencing those manually with automated exports is still the best way to catch anything the scripts might have missed.
