/ˈkɔːpəs/ · Lat. body
1.A word processor, spreadsheets, coding software, an online survey tool and a shared folder.
one platform
for the whole study
corpus is a web platform for sociologists. It lets you run an entire empirical study in one place: participants and consent, interviews and transcripts, field notes, surveys, networks, coding, quotes and data export to R. Every change is logged, raw data is never overwritten, and every figure and quote keeps its provenance. Data is hosted in France.
An empirical study is usually run across several programs that do not talk to each other: transcripts in a word processor, coding in qualitative analysis software, the survey in an online survey tool, response cleaning in spreadsheets, networks in separate graph software, literature in a reference manager, consent forms in a paper binder. What links them together is file names and the researcher’s memory. This leads to four recurring problems.
The interview recording on a voice recorder, the transcript in an email, the consent form on paper, the survey answers in a separate spreadsheet.
In corpus, the interview, its transcript, the survey answers and the consent form are linked to one participant under a code (for example P-014) and open from a single page.
Files called “final”, “final_2”, “edits”. What changed, who changed it and when can no longer be established.
Every processing step on a transcript — cleaning, anonymisation, translation — creates a new version. The original text is kept, and any two versions can be compared word by word.
A supervisor or reviewer asks where the 38% comes from and which questionnaires it was computed on. The answer has to be reconstructed from memory.
Every exported figure links back to the raw answers and cleaning rules, every quote to the transcript version and line.
Participant names in file names, transcripts on a personal laptop or in cloud services hosted outside the EU.
Participant names and contact details are encrypted and never appear in exports. The server is in France.
Screenshots of the platform, version 0.6, on a demo study with fictional data. The qualitative and quantitative data of one study are handled in a single environment. Everything exports to open formats — CSV, JSON, DOCX, R, GraphML — readable without corpus.
Topic, problem, question, hypotheses, methodology, fieldwork, analysis, conclusions. Changes to the methodology are saved as versions with a reason.
A transcript passage gets a code or becomes a quote, with the participant, text version and line.
The “individual — tie type — individual 2” table becomes a graph; density, degree, betweenness and communities are computed.
Survey builder, collection by link, cleaning rules with reasons, export to R with a variable dictionary.
Which interviews and which participant groups each code appears in.
Entry checks catch inconsistencies; a doubtful value becomes a query to whoever entered it.
Who changed what and when, the value before and after. Entries are linked in a hash chain.
← scroll →
ALCOA+ is the data quality standard used in clinical trials. It requires every record to be attributable, legible, contemporaneous, original and accurate, as well as complete, consistent, enduring and available. corpus applies these requirements to sociological data: for every record you can see who entered or changed it, when and why.
ALCOA+ research receipt
sha256 3f9a 07c2 … b81e c21e
intact ✓The data of a sociological study almost always includes personal data. Storage and protection requirements are therefore built into the design of the platform.
§ 01COR-FR
The platform and study data are hosted by a French provider on servers located in France. No data is transferred outside the European Union.
§ 02COR-GDPR
The university or lab remains the data controller; corpus acts as the data processor. A data processing agreement (GDPR art. 28) is being prepared for the pilot.
§ 03COR-PII
Participant names and contact details are encrypted and stored separately from the materials. Consent scans are visible to authorised roles only. Withdrawn consent excludes the participant’s material from all exports, including quotes, statistics and the API.
§ 04COR-LOG
Sign-in with a password and a one-time code sent by email. Roles and permissions are set in each study. Every action is recorded in a hash-chained log.
Empirical studies with qualitative, quantitative or mixed data.
§ 01COR-M1
A master’s dissertation or a PhD: from topic, problem statement and hypotheses to fieldwork and analysis. The supervisor can follow progress and check where every quote comes from.
§ 02COR-LAB
A shared study with roles: lead, researcher, coder, viewer. Coders have no access to participants’ names and contact details in the register; everyone’s actions are in the log.
§ 03COR-MIX
Interviews, a survey and network data from the same field are linked by the participant code — no manual merging of spreadsheets.
§ 04COR-EDU
A demo study with fictional data shows the full cycle, from raw data to a table and a quote. A teaching mode for first-year master’s students is in preparation.
corpus is an independent project developed in Paris by Snezhana Galimova, a second-year master’s student in sociology. Version 0.6 is being prepared for a pilot with research teams. Access is by invitation: write to us and we will open a demo study with fictional data and, for the pilot, a dedicated space for your lab.
corpus status report
Paris · 2026
thank you · nothing lost · nothing rewritten
Articles on working with research data, on how the platform works and on the terms we use.
Step by step: how to record the ties mentioned in interviews in a three-column table, upload it to corpus, choose a layout, compute the measures and interpret them correctly.
5 min read
ALCOA+, audit trails, edit checks, queries, double data entry and database lock: what they are, and why they matter to qualitative and quantitative researchers alike.
6 min read
Which research tasks the platform covers, which programs it replaces, how it handles data and where the data is stored.
5 min read