Stop 01 · Data Science FoundationsTopics · 9/9

Key Tools Overview (Python, SQL, ML, Stats)

In short

The data science toolbox looks huge, but four things carry most of the weight: SQL, Python, statistics and machine learning. Here's what each one actually does for you.

A preview of the toolkit this roadmap builds — Python for analysis and modeling, SQL for pulling data at the source, statistics for rigor, and ML frameworks for building predictive systems.

The mascot tailor with a measuring tape round its neck holds scissors beside a sewing machine.

Think of it like this

A tailor's kit

Scissors cut out just the cloth you need, like SQL pulling the rows you want. The sewing machine does the heavy work, like Python. The measuring tape checks sizes, like statistics. The pattern lets you make the next shirt quickly, like an ML model.

Watch it happen

Where each tool fits

SQL

Python

Stats

1/5SQL pulls just the rows and columns you need from the company database.

The big ideas

What you need to know

SQL is daily bread

Almost every company keeps data in SQL databases or warehouses. You'll write queries nearly every day.

Python is the glue

pandas, NumPy, Matplotlib and scikit-learn cover cleaning, maths, charts and models in one language.

Stats keeps you honest

Averages, spread, sampling and tests stop you from trusting a lucky pattern.

Supporting tools

Jupyter for exploring, Git for saving versions, and a spreadsheet for quick checks.

See it in code

Walk through the code

python

1import sqlite32import pandas as pd3from sklearn.linear_model import LinearRegression4 5conn = sqlite3.connect("shop.db")6df = pd.read_sql("SELECT ad_spend, sales FROM monthly", conn)7 8print(df.describe())   # stats: mean, spread9print(df.corr())       # do they move together?10 11model = LinearRegression()12model.fit(df[["ad_spend"]], df["sales"])13print(model.predict(pd.DataFrame({"ad_spend": [50000]})))

1/3SQL pulls two columns from a local database into a pandas table.

All four tools in one tiny script: SQL to fetch, pandas to hold the table, stats to check, and a model to predict.

Save yourself the pain

Common mistakes

Oops

Learning ten libraries shallowly before mastering SQL and pandas.

Do this instead

Get comfortable with SQL and pandas first. They're used in nearly every task.

Oops

Skipping statistics because libraries 'do the maths'.

Do this instead

Libraries compute numbers. Stats tells you whether those numbers mean anything.

Your mission

Set up your toolkit

~30 min

Quick check

You've found that users who get push notifications order more. Which tool tells you if that difference could just be chance?

Go deeper

1Skim the table of contents of each remaining phase to see how the tools fit together.
2Learn that strong SQL and statistics fundamentals matter as much as flashy ML skills.
3Set a rough personal priority order — most beginners should go math and Python first.