Data Science for Non-Developers

Bring a messy spreadsheet, leave with defensible answers. You will use AI as an analytical partner to clean data, choose the right comparison, test whether a difference is real, build a dashboard people use, and — critically — explain the result without overclaiming.

Data Science for Non-Developers runs to roughly 25 hours of guided video across 8 modules and 40 lessons, taught at intermediate level. Every lesson is recorded as a screen-led demonstration: you watch the work happen, pause, and do the same thing with your own material. Nothing is presented as theory that cannot be performed within a few minutes of hearing about it.

This course sits in The Agentic Workflow, our intermediate track. Intermediate here does not mean programming; it means you will configure systems that act on your behalf, and you will be responsible for what they do. That responsibility is why evaluation, cost control and monitoring appear in every course in this track.

Enrolment is a single payment of £99 and it does not expire. That covers the full video library, the written workbook, the prompt and template pack used in the demonstrations, the student community, and every re-recorded lesson we publish as the tools change. There is no subscription, no upsell ladder and no separate certification fee — when you finish, your certificate is issued automatically.

We wrote Data Science for Non-Developers for a specific person: someone who has read enough about artificial intelligence to be convinced it matters and has not yet found a course that treats them as an adult with limited time. If you have watched hours of free video and still cannot point to something in your week that changed, that is not a failure of intelligence — it is a failure of sequencing, and sequencing is what a structured course provides.

You will be comfortable here if you can use a browser, a spreadsheet and email. Because this is an intermediate course you will also be asked to work inside settings screens, read logs, and connect one service to another using keys and permissions. We show every step, but we do not pretend it is effortless — the reward for that difficulty is a system that keeps working after the lesson ends.

You will not enjoy this course if you are looking for a passive overview of the field, an inspirational lecture series, or a promise of income without work. Every module carries an artefact — a document, a system, a page, a report — and the course is only useful in proportion to how many of those artefacts you actually build.

Students who benefit most tend to share one habit: they bring their own real material to the lessons. Your own inbox, your own product, your own client, your own messy spreadsheet. The techniques transfer, but the confidence only arrives when you have watched them work on something you care about.

Learning outcomes are easy to write and hard to honour, so each of the statements below is tied to a specific lesson and a specific artefact you produce. If you can perform all of them without referring to the videos, you have finished Data Science for Non-Developers, whether or not you watched every minute.

We deliberately keep the list short. A curriculum promising thirty outcomes is usually promising familiarity rather than capability, and familiarity is what free content already provides in abundance. The value of a paid course is the small number of things you can reliably do afterwards.

The course is built in 8 modules, sequenced so that each one depends only on what came before it. You can skip forward, but the order below is the order in which the material compounds, and students who follow it finish materially more often than those who graze.

Each module contains between three and six lessons, a short written summary, the exact prompts and templates used on screen, and a single exercise that produces the artefact for that stage. Exercises are designed to take between fifteen and ninety minutes; anything longer has been split, because the most common reason a course is abandoned is an exercise that felt like a project.

The 25 hours figure is an honest estimate of guided video plus exercises for a student working carefully. Some finish in two focused weekends; others take a term at two lessons an evening. Both patterns work, and neither is penalised.

Tools change quickly, so Data Science for Non-Developers teaches the underlying operation first and the current tool second. When a demonstration goes out of date we re-record it, and your access includes those updates permanently — that is the main practical reason our courses are sold once rather than rented monthly.

The stack used in the demonstrations is listed below. Wherever a free tier can complete the coursework, we use the free tier and say so. Where a paid plan is genuinely required for a specific lesson, we tell you the cost in advance, show what it unlocks, and provide an alternative path so no lesson is gated behind a subscription you did not plan for.

We are deliberately tool-agnostic about the parts that matter. A student who understands why a retrieval step failed can fix it in any product; a student who memorised one interface is helpless the moment a menu moves. Expect to be taught the reasoning, then handed the shortcut.

Every lesson in Data Science for Non-Developers follows the same four-beat structure: the problem in plain terms, the demonstration on screen with nothing edited out, the failure mode you should expect, and the exercise on your own material. The failure mode beat is unusual and it is the one students mention most often — being shown what a bad result looks like is what makes a good result recognisable.

Video is the primary medium because this subject is watched more easily than it is read. You see the actual screen, the actual prompt, the actual moment the model gets it wrong and the correction that fixes it. Alongside the video, each module ships a written companion so you can revisit a step without scrubbing a timeline, and a copyable asset pack so you never retype a template from a video frame.

Assessment is practical and self-paced. At the end of each module you check your artefact against a short rubric — not a quiz on terminology, but a description of what a competent version of that artefact contains. Students who want feedback beyond that post their work in the community, where the same rubric is used for peer review.

There are no live sessions to miss and no cohort to fall behind. That is a deliberate choice: cohort courses have better completion statistics and worse accessibility, and we would rather a shift worker or a parent of young children finished in four months than dropped out in four weeks.

Most organisations are not short of data; they are short of people who will interrogate it. AI removed the technical barrier — a model will write the query, the transformation and the chart — which shifts the scarce skill to framing the question and judging whether the answer holds. That is a reasoning skill, and it is teachable to non-programmers.

The risk grows in proportion to the convenience. A model will happily produce a confident analysis of a badly joined table, and the chart will look just as professional as a correct one. This course therefore drills verification: sanity checks, reconciliation against known totals, and a habit of asking what would have to be true for the result to be wrong.

Career-wise, this is the quiet multiplier. Marketers, operations managers and founders who can answer their own questions in an afternoon make faster, better decisions than those waiting in a queue for an analyst, and they earn credibility that outlasts any particular tool.

Agentic-track graduates are the ones who tend to change roles. Shipping a grounded assistant or an automated process inside an organisation is visible, measurable work, and it frequently converts into a new title, a consulting engagement or a retainer. The skills are also the current bottleneck in the market, which is unusual leverage for a course of this length.

A sensible study path continues sideways rather than upward. Students often pair Data Science for Non-Developers with one course from a different track — a foundations student adding a marketing course, an agentic student adding a masterclass — because capability in this field compounds across domains rather than deepening in one. Every course is the same £99, so combining two is a decision about time rather than budget.

On certification: your certificate of completion names the course, the level, the estimated hours and the date, and it links to a public page describing the curriculum so anyone can see what it involved. We make no accreditation claims. It is a credible record of work done, which is precisely what it should be.

The most common mistake students make is starting on a Sunday with unlimited enthusiasm and no schedule. Data Science for Non-Developers is designed to be completed in fixed blocks, and we recommend booking them the way you would book a class you had paid to attend in person — because you have.

Two sessions of ninety minutes a week finishes this course comfortably inside a couple of months. One session a week finishes it too, more slowly. What consistently fails is watching several modules in a single evening without doing the exercises — the videos feel productive and nothing is retained, because the learning lives in the artefact rather than the viewing.

Keep one document for the whole course. Every artefact, every prompt that worked, every note about something to revisit. By the end it is the most valuable thing you own from the course — more useful than the videos, because it is written in your own context and about your own work.

One payment of £99 for permanent access to Data Science for Non-Developers. We price every course in the college identically, at a level a working adult can decide on without a meeting, because the alternative — tiered pricing designed to make the middle option look sensible — is a sales technique rather than an education policy. Prefer to subscribe? Every course in the college is also available for £19/mo, with the same access and the same updates, cancellable at any time.

If the course is not right for you, write to us within fourteen days and we will refund it. We ask only that you tell us what was missing; that feedback is how the curriculum has been revised each term since the college opened.

A short, honest set of answers. If yours is not here, the admissions team replies to every message — the contact page reaches a person, not a queue.