Data Science Tutorial for Beginners: A 2026 Reality Check

01.The minimum viable stack

Python 3.12, pandas, numpy, scikit-learn, matplotlib and Jupyter. Install via `pip install pandas numpy scikit-learn matplotlib jupyter`. Most companies still use this exact toolset in production notebooks.

02.pandas in one page

A DataFrame is a labelled 2D table. 90% of data work is filtering rows, computing columns and grouping.

import pandas as pd
df = pd.read_csv('orders.csv')
revenue_by_city = (df.query('amount > 1000')
                     .groupby('city')['amount']
                     .sum()
                     .sort_values(ascending=False))
print(revenue_by_city.head())

03.Statistics you'll actually use

Mean vs median, variance, correlation, hypothesis tests. Beyond that, knowing when not to trust a number matters more than knowing fifteen distributions. Always plot before you summarise.

04.Training a first model

scikit-learn's API is the same for every model: `.fit`, `.predict`, `.score`. Start with linear models — they're fast, explainable, and often outperform XGBoost on small datasets.

from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=0.2, random_state=42)
clf = LogisticRegression(max_iter=1000)
clf.fit(X_tr, y_tr)
print(clf.score(X_te, y_te))

05.The professional workflow

Question → data audit → cleaning → exploratory analysis → modelling → presentation. Spend 70% of your time on the first four steps. Modelling is the easy part.