MMT ORBIT ACADEMIA · FOR RESEARCHERS
Your question may be ready while your data is not. Your data may be ready while the right method is unclear. Or you may have the expertise and the means but not yet the question worth pursuing. MMT Orbit Academia starts from where you are.
Academics, researchers, doctoral and master's students running their own work, and research teams. Used with a personal account or from inside your university's organization space.
Equipment, data access, time and expertise are known; the direction must fit them and be worth the work. The output is not a topic list but reasoned directions and the data or experiment each one needs.
Period, variables, data source and a suitable method must be settled together. Accessible data must be separated from the need for a new experiment.
Having a file is not the same as data sufficiency. Columns, units, gaps and measurement conflicts must be visible.
Code or output exists, but which step is missing and which assumption is straining is unclear.
Each objection must connect to its claim, section and data; the need for further analysis must be separated.
MMT investigates what needs investigating, helps gather the sources and data required, prepares method options, runs the supported technical tools and keeps results with their evidence in the same study. Scientific reasoning stays with you; repetitive preparation and tool operation do not.
Searches source candidates, reads what it can retrieve and records verification state. An unread source is never support.
Shows accessible data, data needing preparation and data requiring a new experiment separately; checks files with the verified reader.
Each option carries why it is considered, which assumptions it rests on and which claims it cannot support.
Descriptive statistics, stationarity tests and linear regression run on the same data with the local computation tool. Numbers come from computation, not from a model.
Association versus causation, unmet conditions and limits are recorded together with the output.
When new data or reviewer feedback arrives, the affected decision returns for re-confirmation; old output is never deleted.
Illustrative example; not a real customer study.
Illustrative example artefact names; not taken from a real study.
regression_output.txt
Coefficients, standard errors, diagnostics and unmet conditions; the provenance states which file and period produced them.
data_check_report.md
Columns, units, delimiter, missing periods and measurement conflicts.
source_list.csv
Retrieved sources, retrieval date and reading/verification state.
decision_history
Which method was chosen and why, and which decision was reassessed with new data.
These are shortened examples of the record formats the product really produces. They belong to no real study; the numbers are illustrative.
Question: effect of the minimum wage increase on small-business employment (2015–2024, monthly). Chosen method: difference-in-differences. Control group: the same size class in neighbouring provinces. Reason: the change did not start at the same time everywhere; parallel trends were tested on 2015–2018. Rejected option: a simple before/after comparison — it cannot separate seasonality from the 2020 shock. Decided by: you. The record carries its date and version number.
period,group,employment_index,observations 2019-Q4,treated,100.0,412 2019-Q4,control,100.0,389 2020-Q1,treated,96.4,410
@article{example2023, title={...}, journal={...}, year={2023}, doi={10.xxxx/xxxxx}}
Each entry also records where it came from and when it was read; an unread source is never listed as read.Versions of the same study sit side by side. A new version never overwrites the previous one.
v1 → v2: the method changed
An event study replaced difference-in-differences; separating the 2020 shock was recorded as the reason. The v1 table and charts stay readable.
v2 → v3: the data period grew
2024 data was added; observations went from 4,812 to 5,960 and the tables and charts were regenerated. The old numbers were not deleted.
Unchanged in every version
The question, who decided, and the decision dates. You can compare one version against another.
Scope, period, variable and method choice; interpretation of the result; the answer to counter-explanations; which result stands and the publication decision. MMT proposes, shows its reasons and raises objections; it never decides for you. Ethics, originality and responsibility for new data remain with the researcher.
You do not need to pick a method name first. Describe your question and what you have in your own words.