Welcome! No coding or statistics background needed, just curiosity.
Md Mahbub E Noor · University of Barishal
Using data to answer questions and make better decisions.
Data science lives where all three overlap.
Most data science starts with a table. Here are four text messages, described as data.
| message | length | contains "free" | has a number | label |
|---|---|---|---|---|
| Claim your FREE gift now | 24 | yes | no | spam |
| Class moved to room 305 | 23 | no | yes | ham |
| Free entry: text WIN to 80086 | 29 | yes | yes | spam |
| Ok, see you at the library | 26 | no | no | ham |
a single message, patient or sale
a measured property, used as input (X)
what we want to predict (y)
A model learns how the feature columns relate to the label, then predicts the label for new rows. In Lecture 5 this table has 5,572 rows; in Lecture 6 we call the features X and the label y.
Every tap, purchase and scan leaves a data trail.
Swap guesses for evidence: what to stock, whom to treat first.
Health, finance, farming, research: all need people who read data.
| Traditional software | Data science | |
|---|---|---|
| Example | Calculator, payroll system | Spam filter, face unlock |
| Output | Exact, the same every time | A prediction, with some uncertainty |
| Best when | The rules are known exactly | The rules are too complex to write by hand |
Keywords like "free" catch 4 of 8 messages. Every fix opens a new hole.
A model trained on 30 labeled messages learns its own clues.
Type any message and see which words look spammy to the model.
A loop, not a line: exploring often sends you back to ask a better question. Modeling is one step of six.
What will you watch next? Learns from your history and what similar people enjoyed.
Does this X-ray look healthy? Learns from past scans labeled by doctors.
Is this payment fraud? Learns from normal spending and past fraud cases.
How long will this trip take? Learns from live speeds of phones on the road.
| Lecture | Topic | You will be able to |
|---|---|---|
| 1 | What is data science? | Explain what data science is and how models learn |
| 2.1 to 2.3 | Getting started, Python basics, debugging | Run code offline and online; read and fix errors |
| 3.1 to 3.3 | Decisions, lists and tuples, loops | Build a rule-based spam filter, and see it fail |
| 4.1 to 4.2 | Functions, dictionaries | Count words and score messages from examples |
| 5 | Real data with pandas | Load, clean and explore 5,572 real messages |
| 6 | Build a spam detector | Train, test and honestly evaluate a real model |
| [7 onwards] | [Remaining course topics] | [To be added] |
| Item | Details |
|---|---|
| Classes and labs | [Days, times, rooms]. Labs use Jupyter Notebook on the lab PCs (offline) |
| At home | Google Colab (online), or Jupyter if you install Anaconda on a laptop |
| Assessment | [Homework __%] · [Quizzes __%] · [Midterm __%] · [Project __%] · [Final __%] |
| Homework | After every lecture from Lecture 2. Submit your .ipynb file on [LMS / Google Classroom] |
| Late work | [Late policy] |
| Help | Office hours: [day, time, room] · Contact: [email] |
[For example: asking an AI to explain an error message or a concept, like a tutor.]
[For example: submitting code or answers generated by an AI as your own work.]
You must be able to explain every line you submit. [Say whether and how AI use must be disclosed.]
Office hours: [day, time, room] · Contact: [email]
No question is too basic, today especially.
Watch it again at home to revise.