Datalytech is an analytics practice built on engineering discipline and academic rigour. We take on the work that has to hold up under scrutiny, from a doctoral panel to a board meeting to an external audit.
Datalytech began with a pattern that kept repeating. Researchers arrived holding datasets they had collected carefully but could not analyse with confidence. Organisations arrived holding dashboards nobody trusted, built on numbers nobody could trace back to a source. In both cases the underlying problem was identical: the analysis had been treated as a final formatting step rather than as the part of the work that decides what is true.
We built the practice around the opposite assumption. Method comes first, documentation runs alongside the work rather than after it, and every figure we publish can be traced back to the raw file and the script that produced it. That approach costs a little more time at the start and saves a great deal of it when somebody senior starts asking questions.
From a base in Nairobi we now work with clients across East Africa, Europe and the United States. Some are individual doctoral candidates. Others are utilities, lenders and government agencies running data programmes across several departments. The standard applied to the work does not change between them.
See Our Case StudiesDatalytech is led by the person who does the analysis, which is the single biggest reason the work stays consistent from proposal through to handover.
Founder & Lead Analyst
Datalytech was founded by Godfrey Musyimi Karuku, an electrical engineer who moved into data analysis through the measurement side of the discipline rather than through software. Engineering teaches you to distrust a reading until you know how the instrument produced it, and that instinct now shapes how every dataset entering this practice is treated.
His work sits across three connected areas. As an engineer he understands the systems that generate operational data, from metering and load profiles to sensor logs and plant telemetry. As a data analyst he builds the statistical models, pipelines and dashboards that turn those readings into decisions leadership can defend. As an independent researcher he keeps to academic standards of documentation and reproducibility, which is why methods notes accompany every deliverable rather than arriving only when a reviewer asks.
That combination explains the two kinds of client Datalytech serves best. Researchers get an analyst who reads their methodology chapter properly and can argue for a chosen test in front of a panel. Organisations get someone who has stood next to the equipment producing the numbers and knows where the readings tend to go wrong.
These are the commitments that decide how a project runs here, and they apply to a single chapter of statistical analysis exactly as they apply to a multi department data programme.
Every analysis ships with the reasoning behind it. You get the test we chose, the assumptions we checked, the diagnostics we ran, and the syntax file that produced each table. When a supervisor, auditor or board member asks how a figure came about, the answer is already written down.
Nothing is done by hand in a spreadsheet that cannot be traced later. Cleaning steps, recodes and model specifications live in scripts under version control, so the same raw file produces the same result a year from now, in your hands as easily as in ours.
Technical depth belongs in the appendix, not in the conversation with your stakeholders. Findings are written so a non specialist can follow the argument and act on it, with the full statistical detail sitting underneath for anyone who wants to check the working.
The person who scopes your project is the person who runs the analysis. Work is not passed down to a junior after the proposal is signed, which is why turnaround stays short and context never has to be explained twice.