Data Analysis and Visualisation with R and Python Training Course

Participants can write basic R code using the data structures required for analysis.

• Vectors, matrices, lists and data frames for analytical objects
• Factors and ordered categories for controlled categorical analysis
• Indexing rows, columns and elements with base R syntax
• Functions, arguments and return values for repeated analytical tasks
• Conditional logic using if statements, comparisons and logical operators
• Vectorised calculations replacing row-by-row spreadsheet-style processing

Participants can use Python objects, functions and control structures for data analysis.

• Strings, numbers, lists, dictionaries and tuples for analytical processing
• DataFrames and Series as pandas analytical data structures
• Boolean filtering, indexing and slicing for targeted data selection
• Functions with parameters for reusable transformation and analysis routines
• List comprehensions for concise conditional data preparation tasks
• Exceptions and error messages for diagnosing failed analytical code

Participants can import datasets and assess their suitability for analysis.

• CSV, Excel, JSON and database extracts imported into R
• pandas readers for CSV, Excel, JSON and parquet files
• Data dictionaries defining field names, types, units and permitted values
• Profile missingness, duplicates, ranges and unexpected category values
• Character encoding, date parsing and decimal separator import issues
• Source lineage recorded from original extract through analytical dataset

Participants can clean and reshape analytical data using tidyverse functions.

• dplyr verbs filter, select, mutate, summarise and arrange
• Grouped summaries using group_by and summarise for performance analysis
• tidyr pivot_longer and pivot_wider for reshaping reporting extracts
• String cleaning with stringr patterns, replacement and standardisation
• Date calculations using lubridate periods, intervals and reporting periods
• Missing value treatment using explicit rules and documented assumptions

Participants can transform, aggregate and reshape datasets using pandas.

• DataFrame filtering with loc, query and boolean conditions
• Column creation using assign, arithmetic operations and conditional logic
• Groupby aggregations for counts, totals, rates and averages
• Reshaping data through melt, pivot and pivot_table methods
• Text standardisation using pandas string methods and regular expressions
• Datetime conversion, period extraction and elapsed-time calculations

Participants can combine related datasets while identifying join and reconciliation risks.

• Primary keys, foreign keys and grain in relational datasets
• Inner, left, right and full joins matched to business questions
• Join cardinality checks for one-to-one, one-to-many and many-to-many relationships
• Duplicate keys diagnosed before merging customer, transaction and reference datasets
• Reconciliation totals comparing source records before and after integration
• Lookup tables applied for standardised categories, regions and codes

Participants can identify patterns, anomalies and meaningful segments in analytical data.

• Frequency tables and cross-tabulations for categorical data inspection
• Mean, median, percentiles and dispersion measures for numeric variables
• Distribution shape assessed through histograms, density plots and boxplots
• Outliers investigated using interquartile range and domain thresholds
• Segmentation by customer, product, region, channel and reporting period
• Correlation matrices interpreted without confusing association with causation

Participants can select chart forms and visual encodings for specific analytical questions.

• Question-first chart selection for comparison, composition, distribution and relationship analysis
• Visual encodings using position, length, colour, size and shape
• Axis scaling, baseline choices and log transformations handled transparently
• Colour palettes designed for accessibility, contrast and categorical distinction
• Annotations, reference lines and labels supporting decision interpretation
• Misleading charts identified through distorted scales, aggregation and omitted context

Participants can create polished static charts in R using ggplot2.

• Grammar of graphics layers: data, aesthetics, geoms and scales
• Bar, line, scatter and box plots built with ggplot2
• Faceting comparisons across regions, customer segments and reporting periods
• Statistical annotations using smoothing, confidence intervals and summary layers
• Themes, labels and legends formatted for business reporting standards
• Export settings for presentations, reports and high-resolution publication graphics

Participants can produce analytical visualisations in Python using matplotlib and seaborn.

• matplotlib figure objects, axes and layout controls for reporting charts
• seaborn categorical, distributional and relational plot functions
• Subplots arranged for comparative operational performance reporting
• Confidence intervals and uncertainty bands displayed with suitable chart forms
• Titles, annotations and data labels written for managerial audiences
• Reusable plotting functions created for recurring reporting requirements

Participants can produce interactive and static outputs suited to different audiences.

• Plotly charts with hover information, filters and interactive selection
• Interactive graphics compared with static charts for controlled reporting
• Quarto documents combining narrative, code, tables and visual outputs
• Notebook outputs structured for review, rerunning and audit trails
• Parameters used to generate recurring reports for defined periods
• Output formats selected for HTML, PDF, Word and presentation use

Participants can assess uncertainty and apply basic statistical inference appropriately.

• Population, sample, sampling frame and selection bias distinctions
• Confidence intervals interpreted as ranges around estimated population parameters
• Hypothesis tests framed through null hypotheses, alternatives and significance levels
• T-tests comparing means between independent groups and paired observations
• Chi-square tests examining association between categorical variables
• P-values, effect sizes and practical significance reported together

Participants can build, interpret and assess regression models for numerical outcomes.

• Linear regression links predictors with continuous target variables
• Model formulas specified using business drivers and measurable assumptions
• Coefficient interpretation distinguishes direction, magnitude and reference categories
• Residual plots assess linearity, variance and influential observations
• Multicollinearity checked using correlation patterns and variance inflation factors
• Train-test splits assess out-of-sample predictive performance

Participants can create and evaluate classification models for categorical decisions.

• Logistic regression estimates probabilities for binary outcome classification
• Classification thresholds selected according to decision costs and risk tolerance
• Confusion matrices summarise true positives, false positives and errors
• Precision, recall, specificity and F1 score compared for model evaluation
• ROC curves and area under curve assess ranking performance
• Class imbalance treated through weighting, resampling and threshold adjustment

Participants can analyse temporal patterns and prepare basic forecasts.

• Time indexes, frequency and missing periods validated before forecasting
• Trend, seasonality and irregular variation separated through decomposition
• Rolling averages smooth volatile operational and financial series
• Lagged variables compare current outcomes with prior-period behaviour
• Forecast accuracy measured using MAE, RMSE and MAPE
• Prediction intervals communicate uncertainty around forecast values

Participants can create maps that analyse geographically referenced data responsibly.

• Latitude, longitude and coordinate reference systems for spatial datasets
• Choropleth maps use normalised rates rather than raw counts
• Point maps display service locations, incidents and customer concentrations
• Spatial joins connect observations with administrative boundary datasets
• Map classifications compared using quantiles, equal intervals and natural breaks
• Privacy risks assessed before publishing small-area location data

Participants can document, test and review analytical work for reliable reuse.

• Analytical assumptions documented alongside source data and transformation decisions
• Assertions test row counts, key uniqueness and expected value ranges
• Unit tests verify functions performing recurring calculations correctly
• Peer code review checks readability, logic, dependencies and documentation
• Version control concepts support traceable changes to analytical files
• Publication checklists confirm chart accuracy, disclosure risks and source attribution

Participants can present analytical findings with clear recommendations and limitations.

• Executive summaries state decision question, evidence, recommendation and caveats
• Findings prioritised by materiality, confidence and operational relevance
• Chart narratives explain observed patterns without claiming unsupported causality
• Limitations address data coverage, assumptions, uncertainty and potential bias
• Recommendations link analytical evidence with owners, actions and measures
• Stakeholder questions anticipated through traceable calculations and appendix materials

Participants can build, review and present a complete R or Python analytical report addressing a defined business or public-sector question.

• Capstone question framed with decision owner, scope and success measures
• Raw data profiled, cleaned and integrated through documented transformation steps
• Exploratory analysis identifies patterns, anomalies and priority segments
• Statistical or predictive method selected against the stated question
• Decision-focused visualisations produced with transparent labels and supporting context
• Peer review tests code quality, analytical claims and chart interpretation
• Final presentation communicates recommendations, limitations and next analytical actions