Python for Data Analysis Training Course

Participants can write core Python statements used in data analysis scripts.

• Variables, numeric types, strings and Boolean values in Python
• Lists, tuples, dictionaries and sets for analytical data structures
• Conditional logic using if, elif and else statements
• For loops and list comprehensions for repeated transformations
• Functions with parameters, return values and descriptive docstrings
• Exceptions handled through try, except and explicit error messages

Participants can use NumPy arrays for efficient numerical calculations.

• NumPy arrays, dimensions, shapes and compatible data types
• Vectorised arithmetic replacing row-by-row calculation loops
• Boolean masks for selecting observations meeting defined conditions
• Array indexing and slicing for targeted numerical inspection
• Broadcasting rules applied to rates, adjustments and normalisation
• Missing numerical values handled with NaN-aware NumPy functions

Participants can inspect, select and manipulate pandas DataFrames confidently.

• Series and DataFrames as labelled structures for tabular data
• DataFrame inspection using head, info, describe and sample
• Column selection with brackets, loc and iloc indexers
• Rows filtered through Boolean conditions and compound logical expressions
• Indexes, labels and sorted values managed for reliable selection
• Method chaining for readable multi-step DataFrame transformations

Participants can import common data files and assess their fitness for analysis.

• CSV and Excel imports configured for delimiters, sheets and headers
• Dates, currencies and categorical fields assigned appropriate data types
• Missing-value patterns profiled with counts, percentages and visual checks
• Duplicate records identified using business keys and row comparisons
• Range checks, allowed values and referential checks test data quality
• Data dictionary fields mapped to business definitions and source systems

Participants can clean inconsistent fields and create analysis-ready variables.

• String methods standardise names, codes, whitespace and case
• Regular expressions extract structured values from inconsistent text fields
• Missing values treated through imputation, flags, exclusion and documentation
• Datetime methods derive periods, durations, weekdays and reporting months
• Categorical mappings convert source codes into meaningful analytical groups
• Calculated columns create margins, rates, bands and business indicators

Participants can combine multiple datasets and produce reliable summary tables.

• Merge operations join datasets using validated primary and foreign keys
• Join cardinality checks prevent accidental row multiplication during merges
• Concatenation appends compatible extracts from periods, regions or systems
• Groupby aggregations calculate counts, sums, averages and weighted measures
• Pivot tables summarise measures by segment, period and category
• Melt and pivot reshape wide and long reporting structures

Participants can investigate distributions, outliers and meaningful business segments.

• Frequency tables reveal category concentration and sparse classifications
• Distribution measures include median, quartiles, skewness and interquartile range
• Outliers investigated using IQR rules, z-scores and domain thresholds
• Cross-tabulations compare rates, volumes and values between segments
• Quantile binning creates customer, supplier and performance groups
• Correlation matrices identify candidate relationships requiring further investigation

Participants can create charts that communicate evidence clearly and accurately.

• Matplotlib figures control axes, labels, legends, annotations and layouts
• Seaborn charts show distributions, comparisons, relationships and category patterns
• Chart selection matches analytical questions, measurement scales and audience needs
• Time-series lines distinguish trends, seasonal patterns and reporting periods
• Bar charts ordered by value expose ranking and concentration
• Titles and annotations state findings without overstating analytical certainty

Participants can apply and interpret core statistical methods for business questions.

• Sampling concepts distinguish population measures from sample estimates
• Confidence intervals quantify uncertainty around means, proportions and differences
• Hypothesis tests compare groups using stated null and alternative hypotheses
• T-tests assess mean differences under appropriate distributional assumptions
• Chi-square tests examine association between categorical business variables
• Correlation and simple regression distinguish association from causal claims

Participants can prepare and analyse dated data for trend and forecast assessment.

• Datetime indexes support resampling, rolling calculations and period alignment
• Monthly and weekly resampling aggregates transaction-level time-series records
• Rolling averages smooth volatile measures for operational trend review
• Lagged variables compare current performance against prior periods
• Seasonality inspected through grouped periods, decomposition and visual patterns
• Forecast accuracy assessed using MAE, RMSE and percentage error measures

Participants can extract useful analytical fields from text and JSON data sources.

• JSON structures parsed into dictionaries, lists and normalised DataFrames
• API requests retrieve paginated data with parameters and response checks
• Text cleaning removes punctuation, stop words and inconsistent encoding
• Token counts identify recurring terms within comments and descriptions
• Keyword flags classify records using transparent rule-based logic
• Nested JSON fields flattened while preserving identifiers and relationships

Participants can integrate SQL query outputs into a Python analysis workflow.

• SQL query results loaded into DataFrames for further analysis
• SELECT, WHERE and GROUP BY reduce unnecessary data extraction
• Parameterized queries separate query logic from input values
• Database grain assessed before joining transaction and dimension tables
• Python transformations documented alongside SQL extraction assumptions
• Reconciliation compares extracted totals against source-system control totals

Participants can structure analytical code for reuse, checking and efficient execution.

• Functions isolate repeated cleaning, validation and calculation logic
• Assertions test expected row counts, ranges and permitted categories
• Unit-test principles applied to deterministic analytical functions
• Vectorised pandas operations replace slow row-wise apply patterns
• Memory usage reduced through selected columns and efficient dtypes
• Logging records processing stages, exceptions and data-quality outcomes

Participants can package analytical outputs for review, reuse and controlled handover.

• Executive summaries distinguish findings, evidence, limitations and recommended actions
• Notebook outputs exported as HTML, PDF and reproducible source files
• Data lineage records source files, transformations, owners and refresh dates
• Version control principles protect approved analytical code and documentation
• Personal data minimisation guides field selection and sharing decisions
• Peer review checklists assess logic, readability, evidence and reproducibility

Participants can build, review and present a complete Python analysis notebook addressing a business decision.

• Business question translated into measurable analytical objectives and decision criteria
• Source datasets imported, profiled and documented with explicit quality findings
• Cleaning, joining and feature creation performed through reusable Python code
• Exploratory analysis tests patterns, segments, exceptions and plausible explanations
• Charts and statistics communicate findings with stated assumptions and limitations
• Peer review checks reproducibility, validation evidence and interpretation quality
• Final presentation recommends actions supported by notebook outputs and evidence