Explore an open dataset
Clean a public dataset and write five findings, each backed by one clear chart.
Use statistics, code and machine learning to answer hard questions and guide decisions.
You have reached the end of the beginner stages. Prove what you learned with a project.
You have reached the end of the intermediate stages. Prove what you learned with a project.
You have reached the end of the advanced stages. Prove what you learned with a project.
Clean a public dataset and write five findings, each backed by one clear chart.
Compare two groups in a dataset and explain the result to a non-technical reader.
Train and validate a model on open lending data and explain which factors drive its predictions.
Analyse a sample experiment and recommend whether to ship, with the risks spelled out.
Forecast weekly sales for a product with prediction intervals and a simple baseline to beat.
Publish a small interactive app that lets people try your model and see why it predicts what it does.
The same roadmap as a reading list: every stage, what each topic covers, and which ones are optional.
Distributions, sampling, uncertainty and correlation: the language of data science.
The main language for data work: variables, functions, loops and data structures.
Pull and combine the data you need from databases.
A language built for statistics, strong in research and health sciences.
Load, filter, join, reshape and summarise tables of data.
Missing values, duplicates, wrong types and outliers, handled and documented.
Plot and question the data before modelling it, and write down what surprises you.
Combine code, charts and notes in one reproducible document.
Decide whether a difference is real or noise, and what a p-value does and does not mean.
Linear and logistic regression to explain and predict outcomes.
Plan experiments, choose sample sizes and avoid common traps.
Update beliefs with evidence and express results as probabilities.
Train, tune and compare models with consistent pipelines.
Cross-validation, metrics that fit the question, and checks for leakage.
Turn raw columns into signals a model can use.
Often the strongest choice for tabular business data.
Trends, seasonality and forecasts with honest uncertainty ranges.
Neural networks for images, text and other unstructured data.
Classify, search and summarise text with pretrained models.
Estimate what caused an outcome when you cannot run an experiment.
Lead with the answer, show only the charts that support it, and state the decision you recommend.
Turn an analysis or model into a small app that others can try.
Bias, consent and the harm a wrong prediction can do to real people.
Three or four well-written projects that show your reasoning, not just your code.
They frame questions, gather and clean data, apply statistics and machine learning, and explain results so people can make better decisions.
Analysts focus on describing what happened and reporting it. Data scientists go further into experiments, prediction and modelling. Many data scientists start as analysts.
Many employers like one, but strong projects, solid statistics and clear communication can open the door without it.
© 2026 KamerLink. All rights reserved.