suyogjoshi.com

Software Signal Learning · Stage 2 · Launched

Applied Data Analysis with Python

Learn to clean, explore, visualise, and explain unfamiliar tabular datasets through reproducible workflows and defensible analytical reasoning.

This course focuses on descriptive and exploratory analysis. Predictive-model training, model selection, classification or regression workflows, threshold setting, and machine-learning evaluation belong to Stage 3.

Designed for: learners who can independently write and debug small Python programs using functions, collections, files, and basic error handling.

Format
Live, instructor-led online cohort
Teaching time
14 sessions × 90 minutes
Starting level
Python fundamentals required
Applied outcome
Reproducible real-data investigation
Get notified when applications openCheck your Python readiness

Starting point

Is this course right for you?

A good fit if you…

  • write functions and use lists, dictionaries, and files independently;
  • can trace and fix common errors;
  • want disciplined data investigation rather than model training;
  • value reproducibility and evidence;
  • will practise between sessions.

Probably not the right fit if you…

  • are new to programming;
  • cannot yet organise a small program into functions;
  • want predictive ML, deep learning, or production MLOps;
  • only want a certificate or placement promise.

Prerequisites

  • Variables, conditions, loops, functions, and collections
  • CSV/JSON files and simple exceptions
  • Basic debugging and small multi-function programs
Check your Python readiness

Can you explain control flow, write functions, use record collections, read CSV/JSON, handle a simple exception, debug logic, and organise multiple functions?

What you will be able to do

  • Frame bounded questions and reproducible environments.
  • Use NumPy arrays and vectorised operations.
  • Load and inspect unfamiliar pandas datasets.
  • Select, filter, sort, and derive columns precisely.
  • Diagnose missing, malformed, duplicated, and unsuitable data.
  • Build transparent cleaning workflows.
  • Group, aggregate, pivot, join, concatenate, and reshape data.
  • Work correctly with dates, time, and text.
  • Describe distributions with appropriate measures.
  • Reason about sampling, uncertainty, bias, and comparisons.
  • Investigate relationships without implying causation.
  • Create purposeful, non-misleading visualisations.
  • Communicate findings, caveats, and next questions.
  • Produce a reproducible end-to-end investigation.

These are intended learning capabilities, not guaranteed outcomes independent of prior Python fluency, attendance, practice, submitted work, and individual starting point.

Applied outcome

Investigate and explain a real dataset

Discover and communicate a non-obvious pattern in messy real-world data while showing why the finding may or may not generalise.

  1. Question
  2. Inspect
  3. Diagnose quality
  4. Clean
  5. Explore
  6. Visualise
  7. Check
  8. Explain

Deliverables

  • Reproducible notebook and supporting scripts
  • Environment definition and dataset documentation
  • Processed data and quality/assumption log
  • Visual summary, analytical report, limitations, and next steps

Boundary: this is not a predictive-modelling exercise. Model comparison, thresholds, and ML evaluation belong to Stage 3.

How the course works

Learn live

14 instructor-led online lectures of 90 minutes with theory, live exploration, guided coding, interpretation, and short labs.

Practise and investigate

Maintain a reproducible repository, complete short tasks, and build an end-to-end real-data capstone. Independent effort is additional.

Review and support

Selected work may be reviewed in class or office hours. Exact support windows and recording access will be published before applications open.

Four phases · fourteen lectures

Detailed curriculum

Phase 1 — Establishing an analytical workflow

Move from ordinary Python programs to reproducible analysis and vectorised thinking.

Lecture 1 — From Python programs to analytical workflowsSet up an end-to-end reproducible analysis process.

Workflow and environment

  • Question, acquisition, inspection, cleaning, exploration, communication
  • Virtual environments, packages, notebooks versus scripts, project folders

Reproducibility and practical

  • Requirements, randomness, sources, assumptions
  • Set up the capstone repository and question
Lecture 2 — NumPy and vectorised thinkingUse arrays for efficient numerical operations.

Arrays

  • Creation, shape, dimensions, dtypes, indexing, slicing

Vectorisation

  • Arithmetic, masks, aggregations, broadcasting

Practical

Compare core-Python loops with NumPy for readability and performance.

Capstone contribution

Prepare numerical inspection utilities.

Phase 2 — Working confidently with DataFrames

Inspect, transform, clean, combine, and reshape real tables.

Lecture 3 — pandas Series and DataFramesUnderstand labelled tables and inspect new datasets.

Objects and loading

  • Series, DataFrame, index, columns, dtypes; CSV, JSON, spreadsheet awareness, parsing

Inspection

  • head, tail, sample, info, describe, shape, unique values, memory

Profile the capstone dataset.

Lecture 4 — Selecting, filtering, sorting, and deriving columnsExpress analytical questions as precise DataFrame operations.

Selection and filtering

  • Columns, loc/iloc, labels/positions, masks, conditions, membership, missingness

Transformation

  • Sorting, renaming, derived and conditional columns, chained-assignment caution

Create analysis-ready variables.

Lecture 5 — Data types, missing values, and quality diagnosisDiagnose data quality before analysis.

Types and missingness

  • Numbers as text, dates, categories, nullable types; patterns; drop/retain/impute

Quality profile

  • Duplicates, ranges, categories, summaries, assumption logs

Produce a capstone quality report.

Lecture 6 — Cleaning and transforming messy dataPreserve evidence through transparent cleaning.

Cleaning

  • String normalisation, conversion, mapping, dates, duplicates

Design and audit

  • Raw/interim/processed layers, idempotence, validation, rejected rows, rules, before/after checks
Lecture 7 — Grouping, aggregation, and analytical summariesSummarise segments while recognising hidden risks.

Group operations

  • groupby, multiple/named aggregations, pivot tables, counts, rates, weights

Interpretation

  • Small groups, aggregation bias, Simpson’s paradox, denominators
Lecture 8 — Combining and reshaping datasetsJoin sources correctly and validate relationship structure.

Joins

  • Inner/left/right/outer, keys, cardinality, unmatched rows, duplicate keys, row counts, indicators

Reshaping

  • Wide/long, melt, pivot, concatenation
Lecture 9 — Working with dates, time, and textAnalyse temporal and textual fields correctly.

Dates and time

  • Parsing, components, durations, sorting, resampling awareness, time zones

Text and questions

  • Vectorised strings, patterns, token/length summaries, trends, cohorts, extraction

Phase 3 — Reasoning responsibly from data

Describe distributions, uncertainty, relationships, and visual evidence.

Lecture 10 — Descriptive statistics and distributionsChoose summaries that fit the distribution.

Centre and spread

  • Mean, median, mode, robustness; range, variance, standard deviation, quartiles, IQR

Distributions

  • Histograms, skew, percentiles, outliers, business meaning, units, scale
Lecture 11 — Sampling, uncertainty, and responsible comparisonAvoid overconfident conclusions from observed data.

Samples and uncertainty

  • Methods, selection/non-response/survivorship bias; variability, confidence-interval intuition, practical significance

Comparisons

  • Absolute/relative change, denominators, confounding, causal caution
Lecture 12 — Relationships, correlation, and analytical hypothesesInvestigate association without claiming causation.

Relationships

  • Cross-tabs, group comparisons, scatter plots, covariance, correlation

Hypotheses and cautions

  • Observation/explanation, alternatives, segments, confounding, spurious patterns, leakage preview
Lecture 13 — Data visualisation and evidence-based storytellingCreate clear charts that answer a question.

Selection and implementation

  • Bar, line, histogram, box, scatter; matplotlib, pandas plotting, labels, scales, legends, annotations

Communication

  • Hierarchy, misleading axes, clutter, accessibility, evidence-led narrative

Phase 4 — Complete data investigation

Integrate question framing, quality, cleaning, exploration, visuals, and communication.

Lecture 14 — Capstone: investigate and explain a real datasetCommunicate defensible findings, limitations, and next questions.

Workflow

  • Problem, documentation, quality assessment, cleaning, exploration, visualisation

Deliverables

  • Notebook, processed data, visual summary, analytical report

Outcome

  • Discover a non-obvious pattern, check it multiple ways, explain generalisability

Journey bridge

Move from descriptive to predictive questions in Stage 3.

Time commitment and learner responsibility

Teaching and practice

21 instructor-led hours. Analytical labs, exercises, repository maintenance, and capstone effort are additional; the weekly range is not yet confirmed.

Participation

Arrive with the stated Python foundation. Attendance and practice affect progress; recordings, if offered, do not replace participation.

Technical setup and support

A laptop, Python and notebook environments, editor, and reliable internet are required. Support follows the published cohort policy.

Instructor

Learn with Suyog Joshi

Suyog brings more than 20 years of software engineering, architecture, and enterprise-delivery experience across banking, payments, and complex systems. His teaching emphasises decomposition, debugging, evidence, and maintainable work.

That background supports clear questions, reproducible workflows, explicit assumptions, quality checks, and defensible conclusions.

About Suyog

Fee, application, and cohort process

The configured total fee is ₹8,000 INR, inclusive of applicable taxes. A ₹2,000 deposit is requested only after an offer; the configured remaining balance is ₹6,000.

  1. Review fit, prerequisites, schedule, and policies.
  2. Complete the readiness guidance and submit a free application when open.
  3. The application is reviewed.
  4. Suitable learners receive an offer.
  5. Pay the deposit by the offer deadline.
  6. Pay the balance under the published schedule.
  7. The cohort proceeds when its published conditions are met.
  8. Receive environment-preparation and access instructions.

If the minimum cohort is not met, published cancellation/refund terms apply. Receipt/invoice, certificate criteria, recordings, assignment review, and support details will be finalised before applications open.

Frequently asked questions

How much Python do I need?

You should independently use functions, collections, files, exceptions, and basic debugging.

Is completing Stage 1 mandatory?

No. Direct entry is appropriate when you meet the prerequisites.

Is there a readiness assessment?

The on-page guidance is not an exam and does not collect personal data.

What if I am not ready?

Start with Python Foundations for Data Science.

Does this course teach machine learning?

No. It teaches descriptive and exploratory analysis; predictive modelling is Stage 3.

Why is predictive modelling excluded?

Strong data investigation and communication are distinct capabilities and prerequisites for honest model work.

Which libraries are used?

NumPy, pandas, and matplotlib, with Python’s standard tools where useful.

Notebooks or scripts?

Both: reproducible notebooks with supporting scripts where appropriate.

What datasets are analysed?

Real public or provided tabular datasets; final choices are published with the cohort.

How much statistics is required?

Basic numerical interpretation helps; the course teaches the descriptive and uncertainty concepts it uses.

Are advanced statistical tests covered?

No. The emphasis is responsible descriptive analysis and comparison.

How much practice is expected?

Practice and capstone work are additional to 21 teaching hours; the range will be published before applications open.

Will capstone work receive feedback?

The final review model will be published before applications open.

Are recordings available?

No recording access is promised until the policy is published.

What if I miss a session?

Missed-session arrangements will be published before applications open.

What is required for a certificate?

Criteria are not yet published.

Does applying confirm a seat?

No. A seat requires an offer and applicable payment.

When is payment collected?

Only after an offer; applying is free.

What if the minimum cohort is not reached?

Published cancellation/refund terms govern any payment.

What course should I take next?

Practical Machine Learning Foundations is Stage 3.

Your place in the learning journey

Stage 1 · FoundationsStage 2 · Data analysisStage 3 · Machine learning

After this course, you should be ready to frame prediction problems, establish baselines, and evaluate models in Practical Machine Learning Foundations. Predictive modelling is deliberately deferred until then.

View the complete learning journey

Continue from Python fluency to analytical confidence

Applications are not currently open. Request an update without making a payment or reserving a seat.

Get notified when applications open