Platform/
Privacy Model
Platform, Data Privacy
The AnalyZ Solutions Privacy Model
AnalyZ Solutions processes your data entirely inside your own browser. This guide explains the architecture behind that claim, what it means in practice for each feature, and what stays on your device in every scenario.
6 min read
Platform
Beginner
Researchers and evaluators working with sensitive data — household surveys, clinical records, beneficiary information, interview transcripts — need to know exactly where their data goes when they use an analysis tool. The answer in AnalyZ Solutions is: nowhere. All data processing happens inside your browser, on your own device. This guide explains what that means technically, how it applies to each feature, what is transmitted when you use shared features like dashboards or AI interpretation, and what the architecture means for ethics board and IRB compliance.
The core principle: your raw data — every individual record, every variable, every row of your spreadsheet — is processed entirely inside your browser and never transmitted to any server. This is not a policy choice that could be reversed by a terms-of-service update. It is a consequence of how the platform is built.
01How in-browser processing works
Most web-based analysis platforms work by uploading your file to a server, running the analysis on that server, and returning the results to your browser. AnalyZ Solutions works differently. When you load a file, it is read directly into your browser's memory. The analysis — frequency tables, statistical tests, chart calculations, thematic coding — runs inside your browser using Pyodide, a full Python runtime compiled to WebAssembly that executes client-side.
WebAssembly is a low-level instruction format that modern browsers can run at near-native speed. Pyodide brings the full Python scientific stack — including pandas, NumPy, SciPy, and statsmodels — into that environment. The result is that analyses that would typically require a server backend run entirely within the browser tab you have open, without any data leaving your device.
What this means in concrete terms
- Your file is never uploaded. When you open a CSV, XLSX, SPSS, or Stata file in AnalyZ Solutions, the browser reads it directly from your local disk into memory. The file bytes are not sent anywhere.
- Analysis runs locally. Every calculation — from a simple frequency count to a complex inferential test — executes in your browser. The AnalyZ Solutions servers are not involved in any computation on your data.
- Closing the tab clears the data. Because the data lives in browser memory, it is gone when the session ends. Nothing is retained on any server between sessions.
What the servers do handle. The AnalyZ Solutions servers manage your account, authentication, subscription status, and — when you choose to use them — saved dashboard metadata and canvas content. None of this involves your raw data. The servers also handle the AI interpretation requests described in section 03, but only receive the aggregated chart summaries you explicitly send, not your underlying records.
02Session data: what stays and for how long
By default, your data exists only in memory for the duration of your browser session. When you close the tab, the data is gone. AnalyZ Solutions offers an opt-in local saving feature for users who want to continue working across browser visits without re-loading their file each time.
Default session behaviour
Without local saving enabled, your dataset, all variable labels, coded themes, analysis history, and working notes are held in your browser's in-memory state only. Nothing is written to disk, to the browser's local storage, or to any server. The moment the tab is closed or refreshed, the working state is cleared.
Opt-in local saving
If you turn on local saving in the sidebar, AnalyZ Solutions writes a working copy of your session to your browser's IndexedDB storage — a sandboxed, on-device database that no website other than AnalyZ Solutions can access. This copy is stored on your own device, not on any server, and expires automatically after 24 hours. It is designed for continuity between browser visits, not for long-term storage.
Local saving stores data on your device, not in the cloud. The IndexedDB copy is specific to the browser and device you are using. It is not accessible from another device, another browser, or another user account on the same machine. If you work across multiple devices, you should use the session export feature instead, which saves a portable .analyz.json file to your local disk that you can move and reload manually.
Session export
The session export feature saves your entire working state — data, labels, variable types, analysis history, codes, and themes — as a .analyz.json file to your local disk. This file never touches any server. You can save it to your own secure storage, move it between devices, and reload it in any future session.
03What is transmitted when you use shared features
Three features in AnalyZ Solutions involve a network request: AI interpretation (AnalyZense), dashboard sharing, and Data Canvas publishing. In each case, the data transmitted is carefully scoped to only what is necessary for that feature. Your raw records are not involved in any of them.
| Feature |
What is transmitted |
Raw data sent? |
| AnalyZense interpretation |
The aggregated chart summary visible in the chart itself — counts, percentages, means, or other computed values. No individual records. |
NEVER |
| Dashboard sharing |
The chart-level aggregated data needed to render each chart in the public view, plus any interpretations you have generated. No row-level data. |
NEVER |
| Data Canvas publishing |
The charts, tables, and text blocks you have explicitly placed on the canvas, in their aggregated, display-ready form. No row-level data. |
NEVER |
| Account and authentication |
Email address, subscription status, and session token. No analysis data of any kind. |
NEVER |
| All other analysis features |
Nothing. Statistical tests, frequency tables, qualitative coding, sampling calculations, and all other analysis features run entirely in-browser with no network traffic. |
NEVER |
AnalyZense and AI interpretation in detail
When you click to interpret a chart or generate a Sense Making narrative, AnalyZ Solutions sends the aggregated output of that chart to the AI model — the same numbers already visible to you in the chart, such as the bar heights, category labels, and percentages. The model receives no information about individual respondents, no variable values from rows in your dataset, and no metadata about the file itself. The interpretation is generated from the summary, not the source.
This applies to Interpret All in the dashboard view as well. Each chart's summary is sent individually; the underlying row-level data that produced those summaries remains on your device throughout.
04What this means for sensitive data
Research involving personally identifiable information, clinical data, beneficiary records, or other sensitive data carries specific obligations around how that data is handled. The in-browser architecture of AnalyZ Solutions has direct implications for several of the most common requirements.
IRB and ethics board compliance
Because raw data never leaves the researcher's device and is never stored on any server, the data handling posture of AnalyZ Solutions is straightforward to describe to an ethics board. No data transfer agreement, data processing agreement, or server-side retention policy is required for the analysis itself. We recommend verifying with your specific institution, as requirements vary.
Personally identifiable information
Datasets containing names, ID numbers, addresses, or other identifiers can be analysed without those identifiers ever being transmitted or stored externally. The researcher retains full control of the file at all times. The only data that could appear in a shared dashboard or canvas is what you explicitly choose to include in a chart, always in aggregated form.
Clinical and health research data
Health datasets handled under HIPAA, GDPR health data provisions, or equivalent national frameworks can be analysed in AnalyZ Solutions without triggering the data transfer or third-party processor obligations that apply when data is uploaded to a server. The analysis runs locally; nothing is retained.
Humanitarian and development sector data
Beneficiary data collected in field contexts — displacement surveys, needs assessments, programme monitoring data — often involves commitments to data minimisation and local processing. The in-browser model aligns with these commitments without requiring additional technical controls on the researcher's side.
A note on shared features and sensitive data. If your dataset contains sensitive information, review what your charts actually display before sharing a dashboard or canvas publicly. The raw records are never shared, but a chart showing a breakdown by a sensitive variable — ethnicity, health status, location — may itself be sensitive if the groups are small enough to allow re-identification. This is an analytical judgment that sits with the researcher, not a technical constraint of the platform.
05Summary: what goes where
The table below summarises where each type of data in a typical AnalyZ Solutions session resides and whether it ever leaves your device.
| Data type |
Where it lives |
Leaves your device? |
| Raw dataset (all rows and variables) |
Browser memory only |
NEVER |
| Variable labels and value labels |
Browser memory; optionally IndexedDB (local, 24-hour TTL) |
NEVER |
| Analysis results (tables, test outputs) |
Browser memory only |
NEVER |
| Qualitative codes and themes |
Browser memory; optionally IndexedDB (local, 24-hour TTL) |
NEVER |
| Chart aggregated data (for AI interpretation) |
Transmitted to AI model on request only |
ON REQUEST |
| Dashboard chart data (for sharing) |
Stored on AnalyZ servers when you save a dashboard |
ON REQUEST |
| Session export file (.analyz.json) |
Saved to your local disk by your browser |
NEVER |
| Account information (email, subscription) |
Stored on AnalyZ servers for authentication |
ACCOUNT ONLY |
Frequently Asked Questions
If the data never leaves my browser, how does AnalyZ Solutions run Python analysis?
AnalyZ Solutions uses Pyodide, a port of the CPython interpreter compiled to WebAssembly. This allows a full Python environment — including pandas, NumPy, SciPy, and statsmodels — to run directly inside the browser tab, without any server involvement. The same analysis that would normally require a Python installation or a server backend runs locally in your browser.
Can AnalyZ Solutions employees see my data?
No. Because the data is processed in your browser and never uploaded to any AnalyZ Solutions server, there is no mechanism by which anyone at AnalyZ Solutions could access your raw dataset. What is stored on the servers — dashboard chart summaries, canvas content, and account information — is described in the summary table above.
What happens to my data if I close the browser tab?
If local saving is off (the default), the data in memory is cleared when the tab is closed. Nothing is retained anywhere. If local saving is on, a copy is written to your browser's IndexedDB storage on your device, where it expires after 24 hours. In neither case is anything stored on an external server.
Is AnalyZ Solutions suitable for data subject to GDPR?
The in-browser processing architecture means that personal data is not transferred to a third-party processor for the analysis itself, which simplifies GDPR compliance significantly. When you use dashboard sharing or AI interpretation features, aggregated chart data is transmitted, but this does not involve personal data at the individual level. If you are working under a specific GDPR compliance framework, we recommend reviewing the
Privacy Policy and consulting with your data protection officer, as institutional requirements vary.
What does AnalyZ Solutions do with the chart data sent for AI interpretation?
The aggregated chart summary is sent to the AI model solely to generate the interpretation text. It is not stored, used for training, or associated with your account in any way beyond the current request. The chart data sent is identical to what is already visible in the chart — counts, percentages, category labels — not any individual-level data from your dataset.
Can I use AnalyZ Solutions for research that requires IRB or ethics board approval?
Yes. The in-browser architecture is designed to be compatible with IRB and ethics board requirements around sensitive data handling. Because raw data never leaves the researcher's device and is never stored on any external server, the data handling posture is straightforward to describe and defend. We recommend checking with your specific institution, as board requirements and interpretations vary.
What is the .analyz.json session export file, and is it safe to share?
The session export file is a portable snapshot of your entire working session — your dataset, labels, codes, analysis history, and settings — saved to your local disk in JSON format. It is safe in the sense that it never touches any server; you control it entirely. However, because it contains your full raw dataset, it should be handled with the same care as the original data file. It is not intended for sharing with others; it is a personal backup and continuity mechanism.
Start analysing — your data stays with you.
Free to sign up, free to analyze. No data upload required.
Get started
Related Guides