Data View
Data Management
Descriptive Analysis
Inferential Analysis
Data Visualisation
Scale & Reliability
Sample Size Estimation
Qualitative Analysis
Sense Making
Data Canvas
Session Management
API Integration
Data View

Data View is your first stop after loading a file. Before running any analysis, you need confidence that the data loaded correctly — the right number of rows, the right column names, sensible values in every cell.

It gives you a full-featured preview table with a dual-handle row range slider, automatic missing value detection, value label decoding, and a complete column inventory.

  • Preview any row range — drag the left and right handles to inspect any window of rows (e.g. rows 100–300) in a large dataset
  • Show value labels — toggle decoded labels on/off (e.g. "Male" instead of "1"); auto-enabled for datasets that have value labels
  • Column inventory — expand a full breakdown of every variable: data type, non-null count, missing count, unique values, and a sample value
  • Edit variable labels — add or edit human-readable descriptions for any variable; labels appear throughout the app in tables, charts, and outputs
  • Missing value indicator — green badge if the dataset is complete; orange warning with a count if missing values are present
Why it matters: A 5-minute review in Data View catches import errors, mislabelled columns, and unexpected missing values before they ripple into your analysis. Add variable labels here and every module downstream becomes easier to read and share.
Data Management

Real-world data is rarely clean. Surveys come back with typos, missing answers, inconsistent codes, and rows you didn't expect. Data Management gives you a friendly toolkit to fix these issues before you start analysing.

Every action is reversible — there's an Undo button that keeps your last 10 changes, so you can experiment freely without fear of breaking your data.

  • Create new variables from existing ones (e.g. age groups from age)
  • Recode values — turn "1, 2, 3" into "Low, Medium, High"
  • Rename variables to clearer names
  • Delete variables or rows you don't need
  • Manage duplicates — find and remove repeat entries
  • Split columns — divide a text column on a delimiter into separate columns (colname_A, colname_B, …)
  • Reshape data — convert between wide and long formats for longitudinal or pivot analysis
  • Merge or append two datasets into one
  • Set variable types — tell AnalyZ which columns are categories vs numbers
Why it matters: A clean dataset is the foundation of every reliable analysis. Spend 10 minutes here and save hours of confusion later.
Descriptive Analysis

Descriptive analysis answers basic but essential questions: How many people answered each option? What's the average? Where do most respondents fall? Before you run any fancy statistical test, you need to know what's actually in your data.

This module gives you five ways to summarise: Data Quality (missing values, duplicates, constants, and outliers for every variable), Frequency (counts and percentages), Cross-Tabulate (contingency tables with chi-square), Summary Statistics (mean, median, SD, range, skewness, kurtosis), and Publication Table (grouped characteristics table — the academic "Table 1").

  • Counts and percentages for every category
  • Mean, median, standard deviation, range, quartiles
  • Publication-ready cross-tabs with optional p-values
  • Missing-value summary with patterns highlighted
  • Plain-English interpretation of each result
  • Excel and Word exports for your reports
Why it matters: Understanding the basics of the data is the foundation to tell a great story using advanced analytical approaches.
Inferential Analysis

Descriptive statistics tell you what's in your sample. Inferential statistics let you draw conclusions about the wider population — testing whether observed differences are real, or just chance fluctuations.

Whether you want to compare two groups, examine relationships between variables, or build a model that predicts outcomes, this module has the right tool for the job.

  • t-tests — compare two means (independent, paired, or one-sample)
  • ANOVA — compare three or more group means (one-way and two-way)
  • Chi-square — test associations between categorical variables
  • Correlation — measure how strongly two variables move together
  • Regression — linear, logistic, ordinal, multinomial, and hierarchical
  • Effect sizes automatically calculated (Cohen's d, η², Cramér's V)
  • Plain-English interpretation of every result, ready to copy into your write-up
Why it matters: Every test comes with a clear interpretation in plain English — no need to remember whether a coefficient is positive or negative, or what an odds ratio means.
Data Visualisation

A well-designed chart can communicate something a table never could. This module gives you 15+ chart types — each with smart defaults so you can produce something publishable in seconds, plus a full customisation panel for titles, axis labels, colours, fonts, and backgrounds.

You can build a multi-chart dashboard with multiple charts side-by-side and cross-filtering — clicking any bar, slice, or point dims related data across all other charts simultaneously. Share the dashboard as a live link (no account to view) or download as interactive HTML.

To tell a narrative story with your charts, use Data Canvas — add any chart directly to a slide deck with plain-language headlines and context.

  • Bar chart — counts or means by category
  • Histogram — distribution of a numeric variable
  • Scatter plot — relationship between two numeric variables
  • Box plot — distribution and outliers across groups
  • Line chart — trends over time or sequence
  • Pie / Donut — proportions of a whole
  • Heat map — patterns across two dimensions
  • Violin plot — distribution shape across groups
  • Multi-chart dashboard — assemble multiple charts in a responsive grid
  • Shareable live link — publish with one click; anyone views it without an account
  • Cross-filter support — clicking one chart dims related data across others
  • AnalyZense integration — add AI-generated interpretations per chart
  • Add to Data Canvas — send any chart to a narrative story slide with one click
  • Download as HTML — fully interactive, works offline, no software needed
Why it matters: Every chart is interactive (zoom, hover, pan) and exports to PNG at high resolution. Cross-filtering lets stakeholders explore the data live. Shareable dashboard and Data Canvas links make communicating findings instant — no static screenshots, no PDF attachments, no login required.
Scale & Reliability

If you're using a multi-item questionnaire (like a satisfaction scale, a personality inventory, or a stress measure), you need evidence that the items are actually measuring the same thing. That's what scale validation is for.

This module helps you check internal consistency, identify underlying themes, and reduce many variables into fewer composites — three of the most common steps in psychometric and survey research.

  • Cronbach's Alpha — measures whether items hang together as a scale
  • Reverse-coding helper — automatically flags items pointing the wrong way
  • Item-deleted statistics — see whether removing an item improves the scale
  • Exploratory Factor Analysis (EFA) — discovers the dimensions hiding in your items
  • Principal Component Analysis (PCA) — reduces many variables to fewer composites
  • Scree plots and loadings for interpretation
Why it matters: Reviewers and supervisors expect scale validation evidence in research papers and theses. This module produces exactly what they're looking for.
Sample Size Estimation

One of the most common questions in research is "How big a sample do I need?" Too few participants and your results won't be reliable; too many and you waste time and money.

This module gives you a calculator for the most common study designs. You tell it the size of effect you expect to find and the level of confidence you need; it tells you the minimum number of participants required.

  • Simple random sampling (SRS) — foundational sample size for population surveys
  • Known proportion — when you have a prior estimate to refine your sample size
  • Cluster sampling — when sampling villages, schools, or health facilities
  • Stratified sampling — ensures all subgroups (urban/rural, age bands) are represented
  • Case-control studies — comparing cases with controls in health or social research
  • Cross-sectional surveys — point-in-time population snapshots
  • Longitudinal studies — accounting for attrition across multiple data collection rounds
  • Randomized Controlled Trials (RCTs) — the gold standard for impact evaluation
  • Change in proportion — detecting shifts between two time points or groups
Why it matters: Funding bodies and ethics boards increasingly require a justification for the sample size you propose. This module gives you a defensible answer.
Qualitative Analysis

Numbers only tell part of the story. When you've collected interview transcripts, open-ended survey responses, or free-text comments, you need tools that can handle words rather than numbers.

This module helps you organise text data, identify recurring themes, count word frequencies, and explore how concepts relate to each other — the kind of work that traditionally takes weeks of manual coding.

  • Manage Codes — build your codebook with names, colours, and descriptions
  • Assign Codes — click and drag to highlight text directly in the transcript, then click a code button to assign it instantly; coded passages are highlighted with coloured underlines
  • Review Codes — filter all coded segments by transcript or code, export to Excel
  • Transcript Settings — tag transcripts with metadata (e.g. Gender, Location) for comparative analysis
  • Transcript overview — word count, sentence length, top words, codes used per file
  • Code frequency — bar chart and heatmap of how often each theme appears
  • Word cloud — visual snapshot of the most common words, filterable by code
  • Word search — find every sentence containing a term across all transcripts
  • Code saturation curve — see when new transcripts stop introducing new themes
  • Comparative analysis — compare theme frequency across tagged groups (e.g. rural vs urban)
  • Co-occurrence matrix — which codes appear in the same segment, and how often
  • Code network — visual map of how themes connect, with strength of connection
  • Mutual exclusivity — which codes almost never appear together
  • Code clustering — dendrogram grouping codes into broader thematic families
  • Segment length analysis — which themes generate longer, more elaborated responses
  • Code sequences — Sankey diagram showing how topics flow from one to another
Why it matters: Evaluation researchers and academics working with interview data now have a purpose-built analysis toolkit — no expensive software licence, no setup, no coding skills required. And because it lives alongside the quantitative modules, mixed-methods research no longer requires switching between tools.
Sense Making

You've run your analyses, made your charts, and spotted some patterns. Now what? The hardest part of any analysis project is pulling everything together into a story that makes sense to your audience.

Sense Making is an AI-assisted workspace that helps you synthesise findings across modules — so your final report or presentation flows naturally from the data, not as a disconnected list of test results.

  • Findings synthesis — bring results from different modules into one view
  • Narrative drafting — turn statistical outputs into prose
  • Cross-module insights — spot connections you might have missed
  • Suggested next steps — what to investigate further
  • Export-ready summaries — for reports, papers, and presentations
Why it matters: Most analysis ends with a stack of tables and charts and a struggle to explain what they all mean together. This module helps you bridge that gap.
Data Canvas

Data Canvas turns your analysis outputs into a shareable, scrollable narrative. Add slides from any module (charts, tables, participant quotes, or narrative text), write plain-language insight headlines, and publish as a public link anyone can read in their browser — no account, no software, no design skills needed.

It bridges the gap between statistical output and human communication. Where dashboards enable interactive data exploration, Data Canvas tells the story of what the data means.

  • Chart — any Plotly chart from Visualization or AI-converted from a table
  • Table — frequency tables, crosstabs, test statistics, or any analysis output
  • Quote — coded participant excerpts from qualitative analysis
  • Text — bridging narrative paragraphs you write
  • Draft all headlines — generates plain-language insight headlines for every slide in one call
  • Draft all context — writes audience-facing explanation paragraphs for every slide
  • Convert to chart — AI converts a data table into the most appropriate Plotly chart
  • Improve formatting — refines colours, labels, margins, and layout of any chart
  • Generate infographic — creates a self-contained HTML infographic from table or chart data
  • Refine — give a plain-language instruction to update any chart or infographic
  • Undo — revert any AI transformation to the previous state with one click
  • Public link — publish with one click; anyone reads it without an AnalyZ account
  • Mobile-friendly viewer — scrollable layout with side navigation dots
  • Multiple canvases — create separate stories for different studies or audiences
  • Private by default — canvases stay private until you choose to publish
  • "Add to Canvas" button — present on every analysis output across all modules
Why it matters: Researchers, NGOs, and evaluators need to communicate findings to non-technical audiences — donors, community members, board members. Data Canvas closes the last mile between analysis and communication, replacing static report attachments with live, readable, shareable stories.
Session Management

Analysis projects often span multiple sessions. You clean your data on Monday, run your descriptives on Tuesday, build your model on Wednesday. Without session management, that means re-uploading and re-cleaning every time.

Save your entire workspace — data, variable labels, value labels, action history, and overrides — into a single .analyZ file. Reload it any time to continue exactly where you left off.

  • Your full dataset (with any cleaning steps applied)
  • The original raw dataset (so you can always go back)
  • All variable labels and value labels
  • Variable type overrides (categorical vs continuous)
  • Your last 10 actions for undo support
  • The original filename for reference
Why it matters: Reproducibility is everything in research. A .analyZ file is a portable record of your work that you can share with collaborators or come back to months later.
API Data Integration

Connect AnalyZ directly to a live survey platform — no file download, no manual export. Paste any URL from your survey platform and AnalyZ pulls the data straight into your workspace, complete with variable labels, value codes, and multilingual question text.

Supported platforms include KoBo Toolbox, ODK Central, REDCap, Google Sheets (published as CSV), and any REST endpoint returning JSON or CSV.

  • KoBo Toolbox — kf.kobotoolbox.org and any hosted KoBo instance (NRC, UNHCR, etc.)
  • ODK Central — v2 submissions API
  • REDCap — records API with token authentication
  • Google Sheets — published CSV export URL (no token needed)
  • Custom REST API — any endpoint returning JSON or CSV, with Bearer token, API key, or no auth
  • All survey responses up to the current date (full pagination)
  • Variable labels from the form definition
  • Value labels and answer codes (single and multiple select)
  • Multilingual labels — choose your preferred language in the variable picker
  • Group structure stripped — variables shown by short name, not group/variable path

Before loading, a picker shows all variables with their labels. Search, filter, and select only the variables you need. This keeps your workspace clean and avoids loading system metadata or fields you don't need for your analysis.

Once connected, a ↻ refresh button appears under Connected sources in the sidebar. Click it to pull the latest submissions at any time. Your token is stored in browser session memory only — if you close and reopen the browser, you will be prompted to re-enter it.

Privacy: API tokens are never stored by AnalyZ — only held in browser memory for the duration of your session. Data fetched via API is processed identically to uploaded files and is never written to our servers.