A RESEARCH ASSISTANT FOR DIFFERENCE-IN-DIFFERENCES
Did a policy, program or business change make a difference? Develop the study in Claude or Codex, review the design, and explore the finding in a report that brings the data, estimates and checks together.
Private workspaces · Review before analysis · Invitation-only access
DID-ACTO / FROM THE SHARED SAMPLE LIBRARY
100 ZIP areas · 48 months · one common adoption date
Average effect on monthly median rent. Computed from a simulated practice dataset. Exploratory analysis; the walkthrough explains the comparison and checks.
Follow the sample from data to report →WHAT IT DOES
A change after an intervention can have many explanations. DiD-Acto organizes a difference-in-differences study: comparing changes in exposed units with changes in an appropriate comparison group.
Name the population, outcome, intervention, timing and effect you actually want to understand. The workflow asks for missing information instead of silently filling it in.
Understand the planning process →Callaway–Sant’Anna, Sun–Abraham, BJS imputation, extended TWFE, Gardner did2s, efficient GMM and synthetic methods—matched to the design and effect you want to study.
Explore estimators and tests →Ask how large the effect is, which observations contribute, and whether the checks suggest a different explanation. Revisit those answers in the saved report.
Learn to read the output →BUILT FOR REAL RESEARCH QUESTIONS
Useful questions often start with a policy or operational change that affected some units earlier than others.
Study whether an initiative changed an outcome across regions, institutions or eligible groups. Make rollout timing and other contemporaneous changes explicit.
Explore program evaluation →Investigate phased pricing, service or operational changes across locations. Separate the rollout from common market movements and changing composition.
Explore business analysis →Move from panel data to an approved specification and a durable report. Keep the target effect consistent when comparing eligible estimators.
Explore research workflows →Explore the data and design choices behind each study question.
THE WORKFLOW
Use the portal for files and accounts. Use Claude or Codex to work through the research conversation. DiD-Acto checks and executes the analysis behind the scenes.
Walk through the process →Connect the dataset to your question.
Understand what the data can support.
Make the choices before the fit.
Execute the approved analysis.
Read the evidence in context.
Keep a record you can revisit.
WHAT YOU TAKE AWAY
Your result can include an overall or dynamic effect, uncertainty, support information and checks appropriate to the approved method. The report brings the study question, estimates and diagnostic findings together in one place.
See what the evidence says, where the answer is still uncertain, and which follow-up would be useful.
Explore an example reportRead the exact target, outcome scale and time window.
Inspect uncertainty, comparisons and design limitations.
Return to a saved result and download its report bundle.
A WORKSPACE FOR YOUR TEAM
Uploads and analysis results are private from other analysts. Authorized administrators can review them.
Shared datasets are read-only. Use them to learn the workflow before working with your own study.
Sign in with a dedicated account, change your password, and return to your saved files and reports.
The current service supports files up to 200 MB and a 2 GB workspace per analyst. Analysis jobs may queue.
Files and analysis run on our server. Your AI provider processes the tool results you request through your assistant. Read the privacy and limits guide.
IS IT A FIT?
DiD-Acto is most useful when you can explain what changed, when it changed, who was exposed and why the comparison is credible.
Bring your knowledge of the intervention; DiD-Acto helps translate it into a study you can inspect and explain.
Understand what is supported →YOUR FIRST QUESTION STARTS HERE
Start with a shared sample, learn the workflow, then bring your own data.