Current studies
Data Science
The discipline behind this portfolio: analytics, machine learning and business intelligence.
Portfolio · 2026
Data Science Student
Muhammad AbdalData · Machine Learning · Automation
Turning complex data into clear decisions.
Data Science student in Karachi working across analytics, machine learning, AI automation and business intelligence. I clean, structure and analyse data so the people who rely on it can act with confidence.
01Data Analytics
02Machine Learning
03AI Automation
04Business Intelligence
§ 01Introduction
From the ledger to the dataset.
I came to data science through commerce and financial mathematics, so I start with the question a business is trying to answer.
I'm a Data Science student based in Karachi. Before data science, my studies were in commerce and financial mathematics: how businesses record what happens, and how numbers describe risk, growth and value.
That background shapes how I work with data. A dataset is only useful once it answers something specific, so I care as much about framing the question and preparing the data honestly as I do about the model or chart at the end.
My work sits across four connected areas: analysing and explaining data, building machine learning models, designing Power BI reports for decision-makers, and automating repetitive data work with n8n and AI.
Academic path
- Commerce
- Financial Mathematics
- Data ScienceNow
- 01Data Analytics
- Exploring and explaining data so a defined question gets a clear, defensible answer.
- 02Machine Learning
- Regression, classification and clustering, used where they outperform simpler analysis.
- 03AI Automation
- Language models and n8n workflows that take repetitive data work off people's desks.
- 04Business Intelligence
- Power BI reports structured around the decisions people actually make.
§ 02Selected work
The shape of the work.
Three specimen figures showing how a finished deliverable is structured: a decision report, an automated workflow and a reproducible analysis. They use illustrative shapes, not client data.
Fig. 2.1 · Specimen
Headline indicators first, comparison second, trend third, with filters kept to one side. The reader sees the answer before the detail.
Related service: Power BI Dashboards- Built on a clean data model, not a single flat table
- Every visual tied to a question someone asked
- Filters that narrow the view without hiding context
§ 03Services
Five ways I can help.
Each engagement starts with the decision the work should support, and ends with something a team can keep using without me.
- 01
Data Analysis
Exploratory and descriptive analysis built around a clearly defined question, with findings written in plain language rather than left inside a notebook.
You receive
- Cleaned, documented dataset
- Reproducible analysis notebook
- Written summary of findings
- 02
Making Sense of Complex Data
Untangling messy, inconsistent or multi-source data into a structure that can be trusted, queried and explained.
You receive
- Data quality audit
- Documented cleaning steps
- Data dictionary
- 03
AI Automation
Using language models inside routine processes to classify, summarise, extract and route information, with a person kept in the loop where judgement matters.
You receive
- Scoped automation design
- Prompt and logic documentation
- Handover notes
- 04
n8n Workflows
Connecting the tools a team already uses so data moves between them reliably, without manual copying and pasting.
You receive
- Workflow build
- Error handling and alerts
- Run documentation
- 05
Power BI Dashboards
Reports designed around the decisions people need to make, with a clean data model underneath instead of every available chart on one page.
You receive
- Data model
- Interactive report
- Refresh and maintenance guidance
§ 04Skills & toolkit
Tools, placed where they earn their keep.
Every tool in my kit, mapped to the stage of a data project where it does its work. Hover or focus a stage to trace the pipeline.
Where does the data live, and what can it tell us?
- SQLLanguage
Is it complete, consistent and fit for the question?
- PythonLanguage
- pandasLibrary
- NumPyLibrary
Which patterns hold up, and how well does a model generalise?
- scikit-learnLibrary
- RegressionMethod
- ClassificationMethod
- ClusteringMethod
What does a decision-maker need to see?
- Power BIPlatform
- MatplotlibLibrary
What should happen without anyone pressing a button?
- n8nPlatform
- AI workflowsMethod
§ 05Education
Grounded in numbers, trained in data.
Academic background
Commerce & Financial Mathematics
The business and quantitative grounding I bring to every data problem.
§ 06Method
A method, not a mystery.
The same five steps sit behind every analysis, dashboard and automation, so you always know what happens next and why.
- 1Step 01
Define the question
Agree on the decision the work should support and what a useful answer looks like before touching the data.
- 2Step 02
Audit and clean
Check sources, gaps and inconsistencies, and record every cleaning step so the result can be trusted and repeated.
- 3Step 03
Explore and model
Look for patterns, test them, and only reach for machine learning when it answers the question better than simpler analysis.
- 4Step 04
Communicate clearly
Turn findings into reports, dashboards and short written explanations that people outside data can act on.
- 5Step 05
Automate what repeats
Hand recurring steps to n8n workflows and AI automation, with alerts for when something needs a person.
§ 07Contact
Have data that should be saying more?
Tell me about the question you are trying to answer, the data you already have, or the process you would like to stop doing by hand. LinkedIn is the best place to reach me.
Karachi, Pakistan