Phase 01 · Foundations · python
Python from Zero
- lessons
- 43
- modules
- 11
- stages
- 7
- of learning
- 9h 4m
Why this track exists
A company has data in 100 places: a CSV a vendor emails every morning, a payments API, a database another team owns. Somebody has to fetch it, fix the broken rows, and put it somewhere useful, on a schedule, without being woken up at 3 a.m. every time it fails. Doing that by hand does not scale past a week. Python is how you write it down once and let a machine repeat it.
Most data engineering job posts list Python first. In practice you use a small slice of it constantly: read a file, loop over records, call an API, handle the error, write a log line, exit with the right code. This track teaches that slice properly instead of teaching the whole language badly.
What you need before starting
Nothing about programming
This track starts at what a variable is. If you can use a computer and type, you have enough.
Nothing about data engineering
The first lesson explains what the job actually is before any code appears.
The roadmap
7 stages, in the order they build on each other. Each stage lists the modules and lessons it covers, and what you should be able to do by the end of it.
01Orientation
Before writing code, understand what a data engineer is paid to do and why Python is the tool. Skipping this is why beginners write scripts with no idea what problem they serve.
Getting StartedFree
0/3
What data engineering is, why Python is the language of choice, and how this track works.
By the end of this stage
You can explain in your own words what a pipeline is and what part of it Python handles.
02The language core
Every transformation you will ever write is built from four things: values, decisions, repetition, and functions. Get these solid and the rest of the track is just vocabulary.
FoundationsFree
0/3
Variables, types, data structures, and the building blocks of every Python program.
Module checkpoint
2 conceptual questions to verify mastery
Flow, functions & filesFree preview
0/5
Loops, retries, reusable utilities, and reading warehouse files safely.
By the end of this stage
You can read a CSV off disk, loop through it, clean each row, and write the result back out.
Where people get stuck
Dictionaries and lists of dictionaries are the shape almost all pipeline data takes. If that shape still feels fuzzy, redo the exercises before moving on. Everything later assumes it.
03Real data sources
Files on your laptop are the easy case. Real sources are HTTP APIs that paginate, time zones that lie, and fields that are sometimes null and sometimes the string 'null'.
APIs, streams & objectsFree preview
0/3
Talk to services, stream huge files, and model connectors as classes.
Module checkpoint
1 conceptual questions to verify mastery
Time & validation
0/2
UTC intervals and rejecting dirty rows before they hit the lake.
By the end of this stage
You can pull data from a REST API page by page, parse timestamps correctly, and reject records that fail validation instead of silently corrupting the output.
Where people get stuck
Time zones and naive datetimes cause more production bugs than any other single thing in this stage. Slow down on that lesson.
04Code that survives production
A script that works once is not a pipeline. A pipeline runs unattended, fails safely, retries the transient errors, and does not double-write when it runs twice.
Text & PatternsFree preview
0/2
Parse logs and survive encoding surprises at the ingest boundary.
Reliable Pipelines
0/4
Logging, cleanup, retries, and an honest mental model of concurrency.
- Logging instead of print-debugging12m
- Context managers & resource cleanup10m
- Retries, backoff, and idempotency14m
- Concurrency, asyncio, and the GIL14m
Module checkpoint
1 conceptual questions to verify mastery
Packaging & Config
0/3
12-factor config, pinning, and a CLI entrypoint, including what this browser cannot do.
By the end of this stage
You can write a job with logging, retries with backoff, config separated from code, and idempotent behaviour on rerun.
Where people get stuck
Idempotency is the concept most people nod along to and then get wrong. If you cannot explain what happens when your job runs twice on the same input, you have not got it yet.
05Databases, cloud, and profiling
Pipelines read from databases and write to object storage, and the slow ones get fixed by measuring instead of guessing. This stage covers sqlite3, boto3, and the profiling tools.
Databases, Cloud & Profiling
0/3
Connect to databases with sqlite3, interact with cloud object storage via boto3, and profile performance with timeit and cProfile.
By the end of this stage
You can query a database from Python, move a file to and from cloud storage, and find which line is making a job slow.
06Algorithms for pipelines
Interview rounds and real pipelines both punish quadratic loops. This stage applies complexity, buffers, windowing, and graph ordering to data work.
DSA for Data Engineers
0/13
Core algorithms and data structures applied to data pipelines: complexity, stream buffers, windowing, and graph dependencies.
- Big-O complexity & measurement15m
- Hash maps & frequency counting15m
- Two pointers & in-place array scanning15m
- Sliding window & stream buffers15m
- Prefix sums & range aggregates15m
- Sorting & binary search with bisect15m
- Stacks, queues & monotonic stacks15m
- Heaps & priority queues with heapq15m
- Linked lists & pointer chains15m
- Trees, BST & hierarchical traversals15m
- BFS, DFS & DAG traversals15m
- Union-find & connected components15m
- Dynamic programming fundamentals15m
By the end of this stage
You can say what a piece of code costs as the input grows, and replace a nested loop with a hash lookup or a heap when it matters.
07Tests and the capstone
The difference between a script and something a team will let near production is usually a test suite and a clean entrypoint.
Testing & Capstone
0/2
Prove transforms with tests, then assemble an ingest script that uses the whole track.
By the end of this stage
You finish with an ingest job you can put in a repo and talk about in an interview: it takes a real source, validates, transforms, loads, logs, and has tests.
Practice what Python from Zero teaches
Each card is one idea. Watch it move, read the one-minute explanation, then work through problems from easiest.
Python
Strings
Slice, count and clean text
15 problemsPractice
Python
Dictionaries
Group and count by key
16 problemsPractice
Python
Arrays
Loop, accumulate, reverse
15 problemsPractice
Python
Recursion
A function that calls itself
8 problemsPractice
Python
Hashing
Trade memory for a single pass
13 problemsPractice
Python
Sorting
Order by a key, then reason about neighbours
8 problemsPractice
Finish this track, get a certificate
Complete all 43 lessons and claim a certificate with a link you can add to LinkedIn or send to a recruiter.
Mistakes to avoid on this track
| Common mistake | What to do instead |
|---|---|
| Learning Python from a general tutorial that spends 3 weeks on classes and decorators. | Data engineering Python is mostly files, dicts, requests, and error handling. Learn that first, and go deeper only when a real problem needs it. |
| Reading lessons without running the exercises. | The exercises are where the learning happens. Reading code gives you recognition, writing it gives you recall, and interviews test recall. |
| Catching every exception with a bare except and moving on. | That turns a loud failure into silent data loss. Catch the error you expect, and let the ones you did not expect crash the job so you find out. |
Where to practise this
Where to go after this
Python from Zero reviews & rating
Finished a few lessons? Tell us what worked and what was confusing. Reviews are read by hand and help decide what gets rewritten next.