Python for Research and Analytics Training Course

Participants can write readable Python statements for routine analytical tasks.

• Variables, numeric types and strings used in analytical calculations
• Boolean logic and comparison operators applied to data conditions
• Conditional statements control analytical rules and exception handling
• Loops process repeated tasks while recognising vectorised alternatives
• String formatting produces clear labels, messages and file names
• Built-in functions inspected through signatures, docstrings and help

Participants can organise analytical code using appropriate Python structures and functions.

• Lists, tuples, dictionaries and sets selected for analytical tasks
• List comprehensions create transformed collections from source values
• Functions package repeatable calculations with explicit inputs and outputs
• Default arguments and keyword arguments improve function usability
• Exceptions handled with try, except and informative error messages
• Assertions test assumptions before downstream analytical processing

Participants can import common data formats and record their source characteristics.

• CSV imports specify delimiters, encodings, headers and missing-value markers
• Excel workbooks read from selected sheets, ranges and named columns
• JSON records normalised into tabular structures for analysis
• Date parsing converts source fields into consistent datetime values
• Source files profiled for row counts, fields and data types
• Data dictionaries document variables, units, codes and collection methods

Participants can transform tabular data into an analysis-ready pandas dataset.

• DataFrames inspected using shape, dtypes, head and descriptive summaries
• Column selection and boolean filtering isolate relevant analytical records
• Missing values treated through deletion, imputation and explicit flags
• Categorical variables recoded with mapping dictionaries and ordered categories
• Groupby aggregations calculate counts, rates, totals and summary measures
• Merge operations join datasets using validated key relationships

Participants can test dataset quality and document corrective treatment decisions.

• Completeness checks measure missingness by field, record and subgroup
• Validity rules test ranges, formats, permitted values and dependencies
• Duplicate detection identifies repeated identifiers and near-identical records
• Outlier screening compares distributions using IQR and z-scores
• Reconciliation checks compare totals against authoritative source reports
• Transformation logs record corrections, exclusions and unresolved limitations

Participants can profile distributions, relationships and subgroup patterns before modelling.

• Frequency tables summarise categories, proportions and denominator choices
• Central tendency compared using mean, median and trimmed mean
• Dispersion measured through variance, standard deviation, IQR and percentiles
• Cross-tabulations examine relationships between categorical research variables
• Correlation matrices identify linear associations and potential multicollinearity
• Stratified summaries reveal differences across regions, cohorts or segments

Participants can create charts that accurately communicate analytical findings.

• Chart selection matches comparison, distribution, relationship and time questions
• Bar charts display ranked categories with meaningful scales and labels
• Histograms and boxplots reveal distributions, skewness and outliers
• Scatterplots combine regression lines, grouping and transparency appropriately
• Time-series charts show trends, seasonality and annotated events
• Colour, ordering and annotation improve interpretation without distorting evidence

Participants can retrieve and aggregate relational data using SQL from Python.

• Relational tables interpreted through entities, keys and cardinality relationships
• SELECT, WHERE and ORDER BY retrieve defined analytical populations
• GROUP BY calculations produce counts, sums, averages and rates
• INNER and LEFT JOIN operations combine related source tables
• Common table expressions structure multi-stage analytical queries clearly
• Parameterized queries connect Python notebooks to SQLite databases safely

Participants can translate a research question into measurable variables and analytical choices.

• Research questions framed with population, outcome, comparison and timeframe
• Operational definitions distinguish constructs, indicators, measures and proxies
• Measurement scales determine appropriate summaries, tests and visualisations
• Sampling frames assessed for coverage, selection bias and nonresponse
• Probability and non-probability sampling implications stated accurately
• Confounding variables identified before causal claims are considered

Participants can select, run and interpret common statistical tests in Python.

• Sampling distributions explain standard errors and confidence intervals
• Null and alternative hypotheses specified before statistical testing
• t-tests compare means while checking assumptions and effect sizes
• Chi-square tests assess associations between categorical variables
• p-values interpreted alongside confidence intervals and practical significance
• Multiple-testing risks managed through planned comparisons and corrections

Participants can fit and interpret regression models for explanatory analysis.

• Ordinary least squares models relate continuous outcomes to predictors
• Categorical predictors encoded with reference groups and dummy variables
• Regression coefficients interpreted with units, uncertainty and context
• Residual diagnostics assess linearity, normality and heteroscedasticity assumptions
• Interaction terms test whether effects differ between defined groups
• Model comparison uses adjusted R-squared, AIC and validation evidence

Participants can analyse time-indexed data and produce defensible baseline forecasts.

• Datetime indexes establish regular periods, gaps and aggregation choices
• Trend, seasonality and irregular variation separated through visual inspection
• Rolling means and rolling windows smooth volatile operational measures
• Lag features represent previous-period values and delayed effects
• Train-test splits preserve chronological order for forecast evaluation
• Forecast accuracy measured using MAE, RMSE and MAPE

Participants can retrieve, inspect and store data from documented web APIs.

• API documentation assessed for endpoints, parameters, limits and licences
• HTTP GET requests retrieve structured data with requests
• Query parameters filter API results by geography, date and category
• JSON responses parsed into DataFrames with retained source metadata
• Pagination handled to retrieve complete multi-page result sets
• Status codes, retries and rate limits managed responsibly

Participants can prepare text data and extract basic themes and patterns.

• Text cleaning standardises case, punctuation, whitespace and encoding
• Tokenisation separates documents into analysable words and phrases
• Stop-word lists and stemming choices documented for reproducibility
• Term frequencies identify recurring topics within document collections
• TF-IDF weights distinguish distinctive terms across documents
• Keyword-in-context inspection checks automated findings against source text

Participants can join location data to boundaries and create basic thematic maps.

• Coordinate reference systems selected before distance and area calculations
• GeoJSON and shapefiles loaded into GeoPandas GeoDataFrames
• Spatial joins connect point records with administrative boundary polygons
• Choropleth maps display rates using defensible classification methods
• Geocoding limitations, positional accuracy and privacy risks assessed
• Map legends, scales and sources support accurate spatial interpretation

Participants can build and assess classification models for defined prediction tasks.

• Prediction targets distinguished from explanatory and causal research outcomes
• Train-validation-test partitions prevent optimistic performance estimates
• Logistic regression estimates binary outcome probabilities and classifications
• Decision trees reveal rule-based splits and feature interactions
• Confusion matrices report precision, recall, specificity and accuracy
• Class imbalance addressed through metrics, weights and resampling choices

Participants can prepare an analytical project for review, reuse and responsible handling.

• Project folders separate raw data, processed data, code and outputs
• Git commits record meaningful changes to notebooks and scripts
• README files state purpose, setup steps, sources and limitations
• Requirements files capture package versions for repeatable execution
• Personal data minimisation follows purpose limitation and access controls
• Bias, fairness and disclosure risks assessed before publishing findings

Participants can communicate findings, uncertainty and recommendations to decision-makers.

• Analytical narratives connect question, method, findings and recommended action
• Executive summaries state material results without overstating certainty
• Chart captions explain measures, denominators, periods and source limitations
• Sensitivity analyses show how assumptions affect key conclusions
• Technical appendices document methods, code references and validation checks
• Stakeholder questions anticipated through transparent interpretation and evidence

Participants can build, review and present a complete reproducible Python analysis addressing a defined professional question.

• Problem statement defines decision, population, outcome and success criteria
• Source data imported, profiled and documented within a structured notebook
• Cleaning rules, quality checks and exclusions justified transparently
• Exploratory findings guide statistical, predictive or time-series analysis choices
• Visualisations and tables communicate material findings and uncertainty clearly
• Peer review tests code, claims, limitations and reproducibility evidence