R consulting & statistical modeling | PeekGraph
Home / R
R & statistics

R consulting & statistical modeling

A model you can defend. Not a number you hope.

PeekGraph builds the analysis, the forecast, and the write-up that can be rerun — so the meeting argument is the assumption, not the spreadsheet.

Workstreams

What an R engagement includes

R statistical modeling, forecasting, experiments, Tidyverse, and reproducible reports. Scope is named on the call — nothing is billed from this page.

01 · Models

Statistical modeling

What it is

Regression, mixed models, and the assumptions written down.

The pain

A coefficient in a slide with no interval and no check.

You get

A model, diagnostics, and a sentence leadership can actually use.

02 · Forecast

Forecasting

What it is

A range for next period, not a single heroic point.

The pain

Last year’s actual copied forward, then argued in the room.

You get

A forecast, uncertainty, and a holdout you can show.

03 · Experiments

Tests & experiments

What it is

A/B, pre-post, and the design before more noise is collected.

The pain

A p-value hunting trip after the campaign already ran.

You get

A test plan, a result, and what can (and cannot) be claimed.

04 · Repro

Reproducible reports

What it is

Quarto / R Markdown so the PDF is not a one-off export.

The pain

final_v9.docx and a script that only runs on one laptop.

You get

A document that rebuilds from code when the data file updates.

05 · Tidy

Tidyverse wrangling

What it is

The same grain, named, before the model starts.

The pain

Join errors that silently drop the region the VP cares about.

You get

A pipeline in R that produces the analysis table, not a mystery object.

06 · Handoff

Handoff & review

What it is

Someone else can rerun it. The caveats are in the file.

The pain

The analyst left. The model is a rumor.

You get

A short readme, session info, and the decision the analysis was for.

Fit

Who this is for

Analysts

Already in R. The work still fails a second reader.

Research / eval

A design and a claim that will not melt under questions.

Leadership

A range and a caveat, not a fake-precise point.

Sample

A forecast you can replay

View sample forecast
Sample R forecast and residuals chart
Sample · R

Forecast & residuals

A forecast that cannot be replayed → a model with the residual plot in the pack. Demonstration data, not a client file.

FAQ

R questions before a call

Is this instead of Power BI?

No. R is for the analysis. The dashboard can still be the meeting view.

Python or R?

The stack already in use, unless the problem needs the other one. The call is where that is named.

How much does an R project cost?

A written range comes after the call. Nothing is billed from this page.

Can existing scripts be used?

Yes. Most work starts with the .Rmd or script already on the laptop.

Is confidential company data used as a case study?

No. The sample on this page is a demonstration.

Does R work run remotely with a US team?

Yes. The files stay with you at handoff.