JKUAT / Data Science Club

Start anywhere. Converge on something real.

A semester of data science and machine learning, built week by week. Two sessions every week, beginner-friendly, ending with a project you actually made.

Next session: Tuesday or Thursday, 5.30pm, CLB 106

Join the club
  • Tuesday and Thursday, 5.30pm
  • CLB 106
Loss against training step: the curve falls steeply, then flattens at its minimum.lossstepminimum: converged

Fig. 1 / Every semester starts with high loss. That is the point.

01 / The club

Learn data science by building it, together.

The Data Science Club at Jomo Kenyatta University of Agriculture and Technology meets twice a week in CLB 106. No prior experience needed. Bring a laptop and a question.

  • Open to every JKUAT student

    Any course, any year. If you are curious about data, you belong here.

  • We learn by building

    Each session is hands-on. You write code, run it, break it and fix it with people next to you.

  • Teams from week six

    You form a team and spend the second half of the semester on a project that is yours.

2
Sessions a week
12
Weeks to demo day
None
Experience needed

02 / Semester roadmap

Twelve weeks, drawn as a loss curve.

Early weeks drop fast as you pick up the basics. Later weeks refine. Scroll and the dot descends, lighting each week it reaches, until the curve flattens at demo day.

  1. What we cover

    • Notebooks, variables, lists and functions
    • Reading a CSV file for the first time
    • How the club works and how to get unstuck

    You leave with A working setup and a notebook that loads and prints a real dataset.

    Add to calendar

  2. What we cover

    • DataFrames, filtering and grouping
    • Missing values and messy columns
    • Asking a question and answering it with code

    You leave with A cleaned dataset and three questions you answered from it.

    Add to calendar

  3. What we cover

    • Choosing the right chart for the question
    • Labels, colour and clutter
    • Presenting one finding in one slide

    You leave with A one-page data story you can show to a friend.

    Add to calendar

  4. What we cover

    • What a loss function is
    • Gradient descent, step by step
    • Fitting and plotting a regression with scikit-learn

    You leave with A trained model that predicts a number, and a plot of its loss falling.

    Try it: this is the knob you will tune in week four.

    Fig. 2 / Why the learning rate matters

    step 00 · loss 0.64

    Add to calendar

  5. What we cover

    • Logistic regression
    • Train and test splits
    • Precision, recall and the confusion matrix

    You leave with A classifier and an honest report of how well it really works.

    Add to calendar

  6. What we cover

    • Choosing a dataset and a question
    • Splitting roles and setting milestones
    • A one-page project brief

    You leave with A team, a dataset and a brief that every member has signed off.

    Add to calendar

  7. What we cover

    • Feature engineering basics
    • Overfitting, underfitting and the gap between them
    • Cross-validation and regularisation

    You leave with Your project baseline, with a plan to improve it.

    Add to calendar

  8. What we cover

    • How a decision tree splits data
    • Random forests
    • Reading feature importance

    You leave with A tree-based model that beats your baseline.

    Add to calendar

  9. What we cover

    • Neurons, layers and activation functions
    • Backpropagation without the fear
    • Training a small network in Python

    You leave with A neural network you trained yourself and can explain.

    Add to calendar

  10. What we cover

    • Roles in data, from analyst to ML engineer
    • Building a portfolio that gets read
    • Questions from the room

    You leave with A clear next step for your own path after the semester.

    Add to calendar

  11. What we cover

    • Final model and evaluation
    • Writing up results in plain language
    • Demo rehearsal with feedback

    You leave with A finished project and a three-minute demo ready to present.

    Add to calendar

  12. What we cover

    • Three-minute demos from every team
    • Questions and feedback from the audience
    • Celebrating what the semester produced

    You leave with A project you can put on your CV and a room that watched you build it.

    Add to calendar

Convergence: demo day

03 / What you walk away with

Skills with proof attached.

  • A finished project

    A real model, trained on real data, with a write-up and a demo you can show to employers and lecturers.

  • Working fluency in Python and pandas

    Load, clean, explore and plot data without copying a tutorial line by line.

  • The habit of testing honestly

    You will know how to tell whether a model works, and why a high score can still be wrong.

  • A team and a network

    People across courses and years who you built something with, plus speakers who work in the field.

04 / What you will build

Three starting points for your team project.

From week six your team picks a problem. These are ideas to start the conversation. Bring your own and we will help you scope it.

  1. Predict crop yield from rainfall and soil data

    Combine public weather and soil records to estimate harvests, and find which inputs matter most.

    Regression

  2. Spot spam in SMS and WhatsApp messages

    Train a classifier on short text, then measure the cost of blocking a real message by mistake.

    Classification

  3. Forecast campus footfall through the week

    Use timetable and attendance data with tree-based models to predict when spaces will be busy.

    Forecasting

05 / FAQ

Before you ask.

Do I need to code?
No. We start from the very first line of Python in week one. If you already code, the project weeks give you room to go deeper.
Is it free?
Yes. Membership is free for every JKUAT student. All you need is a laptop; we use free tools that run in the browser if yours is slow.
How do I join?
Fill in the sign-up form at the bottom of this page, then come to CLB 106 on Tuesday or Thursday at 5.30pm. You can join in any week, though earlier is easier.
What do I get at the end?
A project you built with a team, a demo day to present it at, and a portfolio piece you can link on your CV.

06 / Join

The loss only goes down if you show up.

Tell us a little about yourself and come to the next session. You can join in any week.

  • Tuesday and Thursday, 5.30pm
  • CLB 106, JKUAT
  • Open to all JKUAT students, no experience needed

As you would like us to call you.

We only use it to share session updates.

Any course and any year is welcome.

Experience with data or code

Prefer chat?Join the WhatsApp groupplaceholder link, swap before launch

Built with