Data
Data Methodology
On this page we explain where the 16,802 companies in the CompAtlas directory come from, what is left out and how often the data is refreshed.
Updated: October 2026
1. Data sources
The directory is compiled not from a single source but from several open sources that cross-check each other. Closed sources, or those whose terms of use would be violated (e.g. scraping Google Maps), are deliberately excluded.
- Companies' own websites — corporate imprint details from the home and contact pages.
- OpenStreetMap / Overpass — ODbL 1.0, © OpenStreetMap contributors. District centres and business points.
- Overture Maps Places — CDLA-Permissive-2.0, © Overture Maps Foundation. Software/technology categories from the Türkiye extract.
- Technopark and technocity lists — the institutions' own public company directories.
- Google Places API (New) — province/district scanning through the official, paid interface in line with its terms of use.
- Corporate domain lists — only the domain part of an email is used; the addresses themselves are never written anywhere.
2. Web crawling and liveness checks
Before a candidate record enters the directory, its website is visited and checked for the following: does the domain resolve, is the HTTP response successful, is the page a parked/for-sale template, and does the content belong to a company with a Turkish-language or Türkiye-based presence. Records that fail this check are not published.
The liveness check isn't one-off: published records are re-checked regularly. Companies whose sites stop responding permanently are first flagged, then removed from the directory. The record is not deleted, only unpublished — so when the same domain is rediscovered, its history is known.
3. Removing duplicate records
The same company may appear with different spellings in different sources. De-duplication is based on domain normalisation (www, protocol and tracking parameters are stripped; subdomains are reduced to the main domain); name similarity is a secondary signal. Two records with the same domain are never published as two separate companies.
4. Companies without a website
This is the directory's known and accepted limitation: companies without a website are not listed. The reason is that we want liveness to be verifiable — publishing a record that can't be verified lowers trust in the entire directory. Such companies can still enter the directory through the claim flow (corporate email verification).
5. Province and category assignment
The province is derived from address, area code and location data; in case of conflict the record isn't published but sent to a review queue. Service categories are derived from the service descriptions on the company's own site, and a company can be in more than one category. 8,729 companies currently have at least one category assigned. The company owner can always change categories from the dashboard; the owner's choice overrides the automatic assignment.
6. Filtering out personal data
Only corporate, publicly available information is stored. Personal names, personal email addresses and personal phone numbers are filtered at the harvesting stage; text caught by the filter never goes live and is held with a review flag. Details are on the KVKK Privacy Notice and Privacy pages.
7. Verification logic
The verification level shown on a company profile is the sum of independent signals: a live website, a corporate domain, a claimed profile, up-to-date contact details and manual verification. No signal feeds into the ranking as a "score"; it only shows the reader how well the record is verified.
User reviews do not feed into the verification level. Anonymous reviews are unverifiable claims; they produce no rankings, badges or scores.
8. Update frequency
- New company discovery: continuous (crawling pipeline)
- Liveness re-checks: at regular intervals across the whole directory
- Edits to claimed profiles: live immediately
- Numbers on this page: refreshed hourly (snapshot of October 2026)
9. Claiming a company
A company owner claims the profile instantly if their corporate email address is on the same domain as the company's website; otherwise verification is done manually. The company then updates its description, contact details, logo and categories itself — these edits override the automatic compilation. Add / claim a company.
10. Reporting errors and removing data
Every company profile has an "Is something wrong here?" link for reporting a wrong website, a closed company, an incorrect province/category or a duplicate record. Reports go to an admin queue and are reviewed manually. To have a record removed entirely, use the data removal page or write to [email protected].
11. Using the data with attribution
Directory statistics are shared under Creative Commons Attribution 4.0 (CC BY 4.0). When using them, cite the source and the period:
The current figures are available as tables on the Türkiye software sector data page, and in machine-readable form at /api/veri/ozet.
Frequently asked questions
Can I use CompAtlas data with attribution?
Yes. Directory statistics are shared under CC BY 4.0; when using them, simply credit "CompAtlas (compatlas.com.tr)" and state the data period.
Why aren't companies without a website in the directory?
The directory's verification logic relies on website liveness: it's the only signal that lets us show, for free and repeatably, that a record is still active. Companies without a website are therefore out of scope — a known limitation of the directory.
How do I get a wrong record corrected?
Report it through the "Is something wrong here?" link on the company profile. If you own the company, you can claim the profile with your corporate email and update the details directly.