OIS Research Services · Dr Yohanan Ouaknine

Research on sensitive data that never leaves the office

OIS works with government research and strategy units and with academic researchers, from the research question to the report. When the files cannot leave the building, the analysis is written outside on synthetic files and run inside on the real ones.

  • Open-source package for Python and R
  • Nothing to install and no internet on the premises
  • Counts leave the office; records do not

What OIS does

Applied research for organisations whose data must stay where it is

OIS Research Services was founded by Dr Yohanan Ouaknine, formerly Head of the Research Branch of the Israel Prison Service. OIS carries out quantitative and qualitative studies for government bodies, supports researchers and graduate students, and develops open-source tools for analysing confidential administrative data in the place where it is kept.

Research and evaluation

Programme and policy evaluation, statistical and predictive models, qualitative research. From framing the question to a report that can be put in front of decision-makers.

Data that stays put

A way of working in which the confidential files never leave the office. Only an aggregate description goes out, and the analysis is written on synthetic files.

Open-source tools

The oissyntheticdata package, in Python and in R: a single file with no dependencies, open to inspection by any IT or information-security department.

Open source · version 3.0.0

oissyntheticdata: one script, at home and on the premises

One tool in two languages, Python and R, for offices whose confidential data is a folder of Excel and CSV files. It describes the folder in aggregate numbers only, rebuilds synthetic files from that description, and gives your analysis script a small statistics toolbox. You write the script at home on the synthetic files and run it, unchanged, on the real files on the premises.

The flow, in four steps

  1. On the premises: the profile

    Copy one file into the data folder and run it. The tool reads every Excel and CSV file, works out how the tables link to each other, and writes a folder called to_send: JSON files holding counts, shares and rounded quantiles only. The file profile_summary.md lists, column by column, exactly what leaves, so that the person who approves the transfer can read it first.

  2. The office sends the folder

    Only the to_send folder goes out, by ordinary e-mail. It contains no row of the data.

  3. At home: synthetic files and your script

    The same tool rebuilds synthetic files from the JSON: the same file names, sheets, headers and row counts, the same links between tables, and similar distributions. Every value is invented. You write your analysis script on them with the built-in toolbox: statistical tests, regression, interrupted time series, and an HTML report that opens offline.

  4. Back on the premises: the same script, on the real files

    Bring the tool and your script, put them next to the real files, and run. The script does not change; only now are the numbers findings.

Synthetic files are for writing the code and checking that it runs properly. A result computed on them is never a finding: every reported number comes from running the script on the real data, on the premises.

What leaves the office, and what does not

In the data
In the JSON files that leave
In the dataID numbers, names, phone numbers, addresses, free text
In the JSON files that leaveShape only: the length, and whether it is letters or digits
In the dataCategories (district, status, case type)
In the JSON files that leaveLabels and counts; a label seen fewer than five times loses its name
In the dataNumbers
In the JSON files that leaveNine rounded quantiles, the mean and the spread
In the dataDates
In the JSON files that leaveCounts per month and per day of the week; never the date itself
In the dataLinked tables
In the JSON files that leaveHow many rows each key has, and what share of the parent records have any

No row is copied. No count below 5 leaves, and that minimum can be raised.

Install

Python

pip install oissyntheticdata
  • Version 3.0.0 on PyPI, released 12 November 2026
  • A single file, gov_one.py, and the Python standard library only
  • MIT licence

R

install.packages("remotes")
remotes::install_github("yohananouaknine/oissyntheticdata-r")
  • Version 3.0.0 on GitHub
  • Base R only, R 4.2 or later
  • The same JSON files as the Python package: a profile made in R can be turned into synthetic files in Python, and the other way round
  • MIT licence

Nothing is installed on the premises. You carry in two files only: the tool and the script template.

What has been tested, and what has not

  • So far the tools have been tested on invented data only, on Windows, with Python 3.15 and R 4.6.2. They have not yet been run on real ministry files.
  • A pattern that exists in one subgroup only (a relationship between three variables) is not reproduced in the synthetic files.
  • Personal columns are recognised by their header. A column of names under an unusual header has to be marked by hand.
  • Version 3 is a new tool and is not compatible with versions 2.x.

To cite: DOI 10.5281/zenodo.20632932. The Zenodo archive currently holds the versions up to 2.2.0.

Services

What OIS does for you

The service depends on who is asking: a research unit that holds the data, a strategy unit that needs a view ahead, or a researcher who needs access to data and support.

Government research units

You have administrative data and questions, and no way of handing the files to an outside researcher.

  • A full study on data that stays with you. OIS writes the analysis on synthetic files, your staff run it, and only aggregate results leave, after you have approved them.
  • Programme and policy evaluation. Matched comparison groups (PSM, IPTW), interrupted time series, regression and predictive models.
  • Recurring reports. A script written once and run again each year on the updated files.
  • Set-up and training. Your staff learn to produce a profile, to read what leaves and to run scripts, so that you can work the same way with other researchers too.

Strategy and planning units

You need a picture of where things stand and a view ahead, grounded in data and in open sources.

  • Horizon scanning and trend analysis from open sources, on the OFEK platform.
  • Scenario building with research and planning teams.
  • Open-source intelligence (OSINT). Collection and summary, through to a finished report.
  • Forecasts and indicators drawn from the organisation’s own administrative data, as a basis for policy papers.

Academic researchers and research institutes

The grant and the question are yours; OIS provides the infrastructure and the support.

  • Access to confidential administrative data. The profile, the synthetic files, the scripts and a run protocol, so that you keep to your timetable while the data stays with the organisation that holds it.
  • Statistical and methodological advice. Study design, choice of method, analysis and the methods chapter.
  • Support for graduate students working on a master’s thesis or a doctoral dissertation.
  • Qualitative research. Qualitative data analysis (QDA), and simulated interviews for testing an interview guide before fieldwork.
  • Lectures and workshops on research methods, statistics and qualitative analysis.

Two ways of working

Full study

OIS is the principal investigator: from scoping, through the analysis, to the written report.

Infrastructure and support

The study and the grant are yours. OIS supplies the synthetic files, the scripts and the run protocol.

The work is organised around the Five Safes: safe projects, safe people, safe settings, safe data and safe outputs.

Who is behind OIS

Dr Yohanan Ouaknine

Information scientist (PhD, Bar-Ilan University). He served in the Israel Prison Service from 1993 to 2025 and retired as Head of the Research Branch, where he led evaluations of rehabilitation programmes and the development of recidivism-prediction models. He now lectures on research methods, is a research fellow at the Open University of Israel, and is the founder of Yahasra.org.

  • LecturerAriel University, Department of Health Systems Management · Ashkelon Academic College, Department of Criminology
  • Research fellowDHSS Hub, the Open University of Israel: quantitative onomastics and the geography of surnames
  • Formerly Head of the Research BranchIsrael Prison Service, 2022–2025
  • Founder of Yahasra.orgAn online database of the Jewish cemeteries of Morocco: more than 41,000 burial records from 47 cemeteries

Contact

Tell me what you need

A few lines about the question, the kind of data, the constraints and the timetable. I reply within 48 hours.

Please do not send personal data or classified information through this form.

yohanan.ouaknine@ois.co.il +972-50-6279777