Data Analysis

The Data Analysis module in the Chisquares platform is designed to help users uncover trends, make informed decisions, and drive impactful results. This feature provides an intuitive interface for analyzing collected survey data efficiently.

Whether you're conducting academic research, business analytics, or healthcare studies, the Data Analysis module streamlines the process, allowing you to derive meaningful insights.

Features

The Data Analysis module offers powerful functionalities, including:

Data Processing & Analysis

• Import and analyze survey responses with built-in analytical tools.
• Perform statistical operations to identify trends and patterns.
• Generate visualizations for better data interpretation.

Search, Filter & Sorting Options

• Search Bar – Quickly locate past analyses using keywords, names, or IDs.
• Filter – Narrow down results based on specific parameters.
• Sort – Organize analyses based on predefined criteria (e.g., date, relevance).

Automated Insights

• Get AI-driven recommendations based on response patterns.
• Identify outliers and trends with smart analytics.
  Export & Reporting
• Export analysis results in multiple formats (CSV, PDF, etc.).
• Generate summary reports with actionable insights.

Getting Started

To use the Data Analysis module effectively, follow these simple steps:

1. Accessing Data Analysis Module*

• Log in to your Chisquares account.
• In the left-hand navigation bar, select Data Analysis.

2. Launching the Data Analysis Section

Once you click Data Analysis, you will be taken to the Data Analysis page.

Features on This Page:

• Intro text and onboarding link.
• A searchable list of recent analyses.
• A prominent Start Analyzing Now button

3. Creating a New Analysis Project

Click the Start Analyzing Now button to create a new project.

Required Inputs:

• Project Title (mandatory)
• Project Description (optional)

After filling in the details, click Start Analyzing to proceed.

Uploading and Managing Datasets

What Can You Upload?

Chisquares supports the upload of:

• CSV (.csv)

• Excel (.xls, .xlsx)

• Public datasets (cleaned and curated)

• Survey data collected using the platform itself

Each uploaded dataset becomes part of the current project and can be used for immediate analysis.

When to Upload a Dataset

Upload a dataset when:

• You begin a new analysis

• You want to use public data stored by Chisquares

• You’ve collected survey responses through the platform

• You’re replacing an old dataset with a newer version

How to Upload

1. Go to the Project.

2. Click on Upload Dataset.

3. Choose a source:

o Your Device: Upload CSV or Excel

o Your Storage: Use previously uploaded files

o Public Repository: Browse and select from preloaded government/public datasets

o Collected Surveys: Use data gathered through the Chisquares survey engine

4. Validate file:

o The platform checks file type, structure, and size

o File is converted into a secure and efficient format for internal use

5. Review metadata:

o Dataset name and label

o Number of rows/columns, missing values, date imported

Where to Access Uploaded Data

• Go to the Dataset Codebook (automatically loaded post-upload)

Who Can Upload or View Datasets?

• All collaborators in a project can upload and view datasets

Only Project Owners can delete datasets

4. Importing a Dataset

You'll now choose how to import data into your project.

Import Options:

• 📂 Upload datasets (CSV, XLSX, JSON)
• 📋 Import from your survey
• 🤝 Import from peers
• 🌐 Import from other platforms
• 📚 Import from public datasets

Choose the most relevant option for your analysis.

5. Uploading a Dataset

If you select Upload datasets, you’ll see a file picker screen.

Steps:

• Select from *Recent Files, My Storage, or manually *Upload*
• Click to check the dataset
• Click Import

6. Viewing and Managing Your Dataset

Once the dataset is imported:

Key Elements on Screen:

• Dataset information (ID, source, import date, etc.)
• List of variables and their types
• Buttons for Data Processing, Estimate, Test, and Analysis Wizard.
• Prompt to enable Weighted Analysis (optional)

Click Yes to set weights if needed, or No to proceed unweighted.

Step 7: Verify Dataset Information

• Check the Dataset ID and Name at the top left to ensure you're working on the correct dataset.
• Review the Owner and Import Metadata:
  • Imported by: Confirms who uploaded the dataset.
  • Date Imported: Date when data was uploaded.
  • Number of Variables (Columns) and Participants (Rows): Gives a quick overview of dataset size.

Step 8: Use the Codebook Panel

Click the Codebook panel (top right) to:

• View detailed metadata of the selected variable.

• See its distribution and value labels (for categorical).

• Check if the variable has missing values or notes.

  Use this before proceeding with analysis to understand your data.

Exploring the Dataset Codebook

What Is the Codebook?

The Dataset Codebook is your control panel for understanding and managing variables in your uploaded dataset. It provides a live overview of your dataset’s structure and offers quick access to variable-specific actions like renaming, recoding, visualization, and more.

When Do You Use It?

Immediately after uploading a dataset or whenever you want to:

• Review dataset structure

• Check for missing data

• Perform basic descriptive stats

• Rename, recode, or manage variable types

Where to Find It

Once a dataset is uploaded, Chisquares automatically redirects you to the Dataset Codebook page.

How to Use the Codebook

1. View Dataset Overview • Dataset label and actual file name

• Import date and user

• Size: number of rows (observations) and columns (variables)

• Number of missing data

2. Access both Dataset and Variable View Check the box beside the entire dataset, or beside the specific variable you wish to view its attributes.

• Dataset Attributes (summary of all variables)

• Variable Attributes (details of selected variable)

3. Explore Variables

Click on any variable to see:

• Five-number summary (for numeric variables)

• Frequencies (for categorical variables)

• Unique values and missing counts

4. Edit or Transform Variables

Next to each variable, you can:

• Rename or Relabel

• Recode values

• Respecify Type (e.g., from integer to categorical)

• Clone a variable

• Delete (with recovery option)

• Visualize: plot a graph for the variable

• Cross-tabulate with another categorical variable

5. Add Notes and Tags

Tagging helps organize variables:

• Add project-specific notes

• Use searchable tags for easier variable management

6. Review Analysis History

Track all actions taken on the dataset:

• Transformations

• Analyses

• Recode operations

7. Monitor Dataset Modifications

A “Last Modified” tag at the top shows the date and time of the most recent change.

Who Can Use the Codebook?

• All collaborators can view and interact with the codebook

• Only users with edit privileges can modify variable settings or delete variables

Why Use the Codebook?

• Gain rapid insight into your dataset’s health and readiness

• Efficiently prepare data for analysis

• Collaborate seamlessly with full traceability of changes

Working with Variables (Rename, Recode, Clone, Delete)

What Is Variable Management?

Variable management allows you to refine your dataset by renaming labels, reclassifying types, cleaning values, and performing basic transformations — all without writing code.

When to Use These Tools

Use these features when:

• Preparing your data for analysis

• Cleaning inconsistent or unclear variable names

• Adjusting variable types (e.g., converting integer to categorical, such as a socioeconomic status indicator coded as “1”, “2”, “3” which should be treated as categorical, not numeric)

• Creating working copies or simplified versions of variables

Where to Find These Options

In the Dataset Codebook or Variable Navigation Bar, hover over or click the three-dot icon beside any variable to access the following actions:

Actions Menu → Rename / Recode / Clone / Delete / Respecify Type

1. Rename Variable Label

• Purpose: Make variables easier to understand without affecting the original data name.

• Use Case: Rename “v1_age” to “age_at_enrollment”

• Steps:

o Click the three-dot menu → Rename

o Enter a new label (max 20 characters)

o Save

• Validation:

o Label must be unique

o Must use letters, numbers, or underscores only

2. Recode Variable Values

• Purpose: Change values or categories (e.g., merge “Male” and “Other” into one group)

• Use Case: Combine similar responses into fewer categories

• Modes:

o Visual (drag-and-drop interface)

o Classic (condition-based logic)

• Features:

o Create new variable or replace original

o Treat selected values as missing

o Recover deleted or merged categories

3. Clone Variable

• Purpose: Create an exact copy of a variable to test transformations

• Use Case: Clone “age” as “age_grouped” for categorization

• Steps:

1. Click the three-dot menu → Clone

2. Provide a unique name for the new variable

3. Save

4. Delete Variable

• Purpose: Remove an unwanted variable from your dataset

• Use Case: Remove a variable that was erroneously created to avoid confusion. • Steps:

1. Click the three-dot menu → Delete

2. Confirm deletion

3. Restore from history using “Recover” option

5. Respecify Variable Type

• Purpose: Adjust how a variable is treated during analysis (e.g., integer vs. categorical)

• Use Case: Reclassify a scale score as categorical for group comparisons

• Steps:

o Click the three-dot menu → Respecify Type. You can also reclassify variables from the arrow beside each variable’s classification on the codebook

o Select new type (numeric, categorical, date, string)

o Confirm

• Validations:

o Note that the system automatically ensures that each type switch is compatible with data format.

o E.g., cannot convert string to numeric unless values are strictly numbers

Why This Matters

Clean, accurate, and well-structured variables: • Improve analysis quality

• Reduce manual errors

Visualizing Variables and Creating Charts

What Is Visualization in Chisquares?

Visualization lets you convert variable distributions into clear, customizable graphics. It helps explore data trends, identify outliers, and generate publication-ready figures.

When to Visualize Variables

Use visualization to:

• Understand variable distributions

• Compare groups and categories

• Create manuscript-ready figures

• Explore relationships between variables

Where to Find It

You can visualize any variable by:

1. Going to the Dataset Codebook or Variable Navigation Bar

2. Clicking the visualize icon next to a variable

3. Once the visualization is shown, you can change the type into any of the alternatives shown.

How Visualization Works

1. Default Visualization

• When you click “Visualize,” Chisquares creates a basic histogram by default.

• Missing values are excluded. • Visualization is done with unweighted data.

Why Visualize?

• Improve understanding of data distribution and spread

• Identify errors or outliers before analysis

• Generate figures for publication or peer review

• Save time on formatting and exporting visuals

Transforming Data – From Simple Calculations to Complex Recoding

What Is Data Transformation?

Data transformation allows you to modify existing variables or create new ones from them — whether by performing arithmetic, changing formats, or grouping values. It’s the bridge between raw data and insightful analysis.

When to Use Transformations

Use transformation tools when you need to:

• Convert continuous data into categories (e.g., age groups)

• Calculate new variables (e.g., BMI, income-to-poverty ratio)

• Extract information from strings or dates

• Standardize formatting for analysis

Where to Access Transformation Tools

There are two main places to initiate transformations:

• From the three-dot menu next to a variable in the Variable Navigation Bar

• From the Data processing panel in the Dataset Codebook

Types of Transformations Available

🔹 Single Variable Transformations

• Numeric Adjustments: log, sqrt, exponent, cube, reciprocal

• Scaling and Standardizing

• Extract Date Components: Year, Month, Day

• Calculate Date Differences (vs. a fixed date)

• String Split by Delimiter

• Arithmetic with Constants: Multiply/divide/add/subtract a constant

🔹 Recoding Variables

• Visual Recode (Drag & Drop):

o Reclassify categorical or continuous variables

o Merge, rename, isolate or group values

• Classic Recode (Rules-based):

o Create rules using logical operators (e.g., Age ≥ 18 AND < 35)

🔹 Composite Variable Generation • Boolean Logic Builder:

o Create categories based on multiple variables (e.g., Male AND Smoker)

• Mathematical Functions:

o Add, subtract, multiply or divide multiple variables

• Date Interval Calculations:

o Number of days/months/years between two date variables

• Unique Combinations:

o Cross-classify two categorical variables

How to Perform a Transformation (Example: Recoding Age)

1. Click the three-dot icon beside the variable “Age”

2. Choose Recode → Visual Method

3. Drag 0–12 into a group and label “Preteens”

4. Drag 13–19 into a group and label “Teens”

5. Continue for “Young Adults”, “Middle-Aged”, “Older Adults”

6. Choose whether to:

o Replace the original variable OR

o Create a new one (e.g., age_grouped)

7. Click Confirm

Who Can Use Transformation Tools?

All project collaborators have access to transformation tools, but only those with edit permissions can apply irreversible changes like deletion.

Why It Matters

Transforming your data prepares it for meaningful analysis. These tools are designed for:

• Flexibility (visual or logic-based options)

• Transparency (actions tracked in history)

• Speed (no-code interface with point-and-click)

Running Your First Analysis with the Estimate Section

What Is the Estimate Section?

The Estimate section is where you generate tables and figures summarizing your data. This includes population characteristics, mean and prevalence estimates, trends, and regression models. It supports seamless manuscript generation by allowing you to insert outputs directly into your document — with explanatory text.

When to Use It

Use the Estimate section when you're ready to:

• Describe your sample (e.g., age, sex, education)

• Report descriptive statistics (means, prevalence)

• Examine trends over time or across groups

• Run regressions (logistic, linear, etc.)

• Populate your manuscript with text and figures

Where to Find It

Navigate to:

Analysis → Estimate Section

Then select one of the available modules:

• Population Characteristics

• Mean Estimates

• Prevalence Estimates

• Trend Analysis

• Regression Models

• Analysis Wizard (guided setup)

How to Run an Estimate

Step 1: Select the Population

• Analyze the entire dataset (default), or

• Define a subset using filters (e.g., Age ≥ 18, Gender = Female)

• A live count of matching rows will appear

Step 2: Choose Your Variables

• Select one or more outcome variables (numeric or categorical)

• For stratified analysis, optionally choose grouping variables

• Specify variable types or reclassify if needed

Step 3: Customize Output Settings

• Choose confidence intervals or standard deviations

• Toggle inclusion of missing values

• Set cutoff thresholds for precision (e.g., coefficient of variation)

Step 4: Click “Analyze”

• Chisquares generates a modal popup with:

o Table of estimates

o Figure (auto-generated chart)

o Metadata (population subset, weights, method, etc.)

Step 5: Push to Manuscript

Choose one of the following:

• Push Table to Manuscript (adds with text summary)

• Push Figure to Manuscript (adds with caption)

• Push Both

• Download results or share via link/email

Why Use It?

The Estimate section:

• Produces manuscript-ready outputs

• Includes structured text based on results

• Automatically tracks source variables, filters, and settings

• Saves hours on formatting and copy/paste tasks

Using the Regression Analysis Module

What Is Regression Analysis?

Regression analysis allows you to examine the relationship between one or more independent variables and a dependent (outcome) variable. It is essential for uncovering trends, associations, and predictive relationships in your data.

When to Use It

Use regression when you need to:

• Adjust for confounding variables

• Predict outcomes based on multiple inputs

• Estimate associations between variables (e.g., odds ratios)

• Test hypotheses about linear or non-linear relationships

Where to Find It

Navigate to:

Analysis → Estimate Section → Regression Models

You’ll find options for:

• Linear Regression

• Binary Logistic Regression

• Probit Regression

• Ordinal Logistic Regression

• Multinomial Logistic Regression

• Poisson and Negative Binomial Models

How to Set Up a Regression Model

Step 1: Choose subsetting of the population

• You may run models on subsets of your dataset or on the whole dataset

If using a subset of the population, define subset criteria using filters and operators (e.g., Age ≥ 18)

Step 2: Choose Outcome Variable

• Must be numeric or categorical, depending on model type

• The platform filters variable types based on selected model

Step 3: Choose Predictor Variables

• Use multi-select dropdown to choose independent variables

• Recode or respecify variables if necessary

Step 4: Specify Settings

• Choose reference groups for categorical predictors

• Select data weighting (if applicable)

Step 5: Run the Model

• Click Analyze

• Chisquares displays:

o Regression coefficients

o 95% Confidence Intervals

o p-values

o Goodness-of-fit indicators (R², pseudo R², etc.)

o Metadata and model diagnostics

Step 6: Push to Manuscript

• Add results as:

o Table only

o Figure only

o Both table and figure

o Full text interpretation

Why Use This Module?

The regression module:

• Ensures model validity with smart defaults and validations

• Generates formatted outputs and interpretations

• Reduces risk of coding errors and inconsistencies

• Supports transparent and reproducible modeling

Creating and Using Tables for Prevalence and Means

What Are Prevalence and Mean Tables?

These are structured summary tables that show either: • Prevalence: Percent of the population with a given characteristic

• Mean: Average value of a numeric variable

These outputs are manuscript-ready and can be stratified by one or more grouping variables. The tables are designed to mirror what’s expected in scientific publications.

When to Use This Feature

Use when you need to:

• Report percentages or averages

• Compare metrics across subgroups (e.g., age, gender)

• Highlight public health trends or disparities

• Generate tables with text for the Results section

Where to Find It

Go to:

Analysis → Estimate Section → Prevalence Estimates or Mean Estimates

How to Create a Table

Step 1: Define Population

• Default: Analyze the entire dataset

• Optional: Define a subpopulation using filters (e.g., Gender = Female, Age ≥ 50)

• Live preview shows number of eligible rows

Step 2: Choose Outcome Variables

• For Mean Tables: Only numeric variables are allowed

• For Prevalence Tables: Only categorical variables are allowed

• Use “Show Eligible” button to filter based on variable type

Step 3: Choose Stratification Variables (Optional)

• Select one or more categorical variables to break down results by group

• Automatically generates separate results for each subgroup

Step 4: Customize Display

• Choose output style:

o Means with standard deviations

o Means with 95% confidence intervals

• Adjust Coefficient of Variation cutoff to suppress imprecise estimates

Step 5: Click “Analyze”

Chisquares generates:

• A table with your results

• A figure for visual summary

• Metadata on filters, stratification, and weights used

What You Can Do With the Output

• Push Table to Manuscript: Inserts formatted table and descriptive text

• Push Figure to Manuscript: Inserts graph only

• Push Both: Complete results + graphics

• Download Table: Save locally as CSV or Excel

• Share Results: Send link to collaborators

Why Use This?

• Automates formatting for publication

• Built-in logic ensures accurate stratification and interpretation

• Saves hours on manual calculations and chart building

Performing Trend Analyses Across Time or Groups

What Is Trend Analysis?

Trend analysis allows you to evaluate how a variable changes across time or across categories such as geographic regions or demographic groups. It is especially useful in epidemiological and public health studies to detect patterns and shifts.

When to Use Trend Analysis

Use trend analysis to:

• Detect increases or decreases in prevalence or means

• Identify seasonal or regional variations

• Evaluate intervention effects over time

Where to Access It

The trend analysis feature is not a stand-alone feature but is nested within Mean estimate or Prevalence estimate set-ups. There is a checkbox, which when enabled, allows the user to supply the time variable for trend, which must be an integer variable.

To access trend analysis, navigate to:

Analysis → Estimate Section →Mean/Prevalence Estimates Trend Analysis checkbox

How to Run a Trend Analysis

Step 1: Select the Population

• Analyze full dataset or define a subset (e.g., Females age ≥ 50)

• Use filters to define subgroups

• System displays eligible rows in real-time

Step 2: Choose the Outcome Variable

• Select a categorical or numeric variable whose trend you want to assess

• Examples: smoking status, income level, BMI

Step 3: Choose the Trend Variable

• Select the variable that defines time or group order

• Must be ordinal (e.g., Year, Age Group, Region Rank)

• Use respecify tool if your trend variable is misclassified

Step 4: Optional – Stratify by Additional Variables

• Add a secondary grouping variable (e.g., Gender)

• Enables side-by-side trend comparison

Step 5: Run the Analysis

• Click Analyze

• Chisquares generates:

o A line chart visualizing the trend

o A trend table with summary statistics

o Statistical output including p-values for linear trend (if applicable)

o Metadata about dataset, filters, and groupings

Step 6: Export or Insert Into Manuscript

• Push figure, table, or both to the manuscript with accompanying text

• Download as image/CSV or share a link

Why Use This?

• Visually summarize changes over time or ordered categories

• Automatically test for statistical significance of trends

• Quickly populate your manuscript with interpretable figures

• Saves time generating graphs and formatting text

Using the Analysis Wizard (for Beginners)

What Is the Analysis Wizard?

The Analysis Wizard is a guided, step-by-step assistant that helps new or non-technical users set up their analyses quickly and accurately. It simplifies the decision-making process by suggesting options and validating choices at each stage.

When to Use It

Use the wizard if you:

• Are unfamiliar with statistical methods

• Want help selecting variables, models, or settings

• Prefer a conversational setup process

Where to Find It

Navigate to:

Analysis → Analysis Wizard

How the Wizard Works

Step 1: Choose One Main/Primary Outcome (1 single variable)

• Choose the main outcome variable

• Specify or confirm its type

Step 2: Choose 1-5 Secondary Outcomes

• Choose your secondary outcome variables (numeric or categorical) • Specify or confirm their type

• The wizard checks variable type and compatibility

Step 3: Select Main Predictor or Exposure Variable

• Choose the main predictor variable

• Specify or confirm its type

Step 4: Choose the Demographic or other key variables (e.g., confounders)

• Choose independent variables for regression

• Select categorical variables to stratify outputs

• Wizard filters available options by type

Step 5: Review Summary

• The Wizard displays your selections:

o Population characteristics

o Outcome and predictors

o Type of analysis suggested

Step 6: Run Analysis

• Click Continue to launch the appropriate module (e.g., regression, prevalence, trend)

• Output is generated as usual with table, figure, and metadata

Step 7: Push Results to Manuscript

• Choose to push outputs directly into your manuscript

• Download, export, or share with your collaborators

Why Use the Wizard?

• Speeds up analysis setup

• Removes guesswork for beginners

• Prevents invalid selections and incompatible inputs

• Provides explanations and examples along the way

Exporting, Sharing, and Saving Your Results

What Can You Export or Share?

Once your analysis is complete, you can export or share: • Tables (CSV, Excel, or HTML)

• Figures (PNG or SVG)

• Auto-generated text (editable)

• Full datasets or transformed versions

When to Export or Share

• Finalizing your manuscript

• Collaborating with co-authors or reviewers

• Presenting your results externally

• Backing up key outputs

Where to Access Export and Sharing Options

From any modal or result popup:

Click Download, Share, or Push to Manuscript buttons at the bottom You can also access:

• Data Export tools from the Dataset Codebook

Exporting Results

Tables

• Export as CSV, Excel, or HTML

• Automatically formatted for manuscript or submission

Figures

• Download as PNG (for presentations)

Auto-Generated Text

• Text pushed to the manuscript can be copied, exported, or edited directly

Step 9: Prepare Your Data

Click the Data Processing dropdown for options such as:

• Clean Data: Handle missing values, remove duplicates.
• Recode Variables: Combine or modify values in categorical variables.
• Create Derived Variables: Generate new variables from existing ones.

Ensure your dataset is tidy and formatted before running tests.

Step 10: Choose Analysis Type

You have three main options:

1. Estimate
Use this when you want to:

• Compute descriptive statistics (e.g., mean, median, proportion).
• Estimate confidence intervals.

2. Test

• Use this to perform parametric and non-parametric statistical tests.
• Select your test from the Test dropdown menu.

3. Analysis Wizard
For guided analysis:

• Click Analysis Wizard to launch a step-by-step assistant.
• Ideal for non-technical users or those unfamiliar with stats.

Additional Features

Actions Menu (for Each Variable) Each variable has an Actions column with icons for:

• Viewing variable distribution.
• Editing metadata or applying transformations.
• Tagging for quick filtering.

Weights & Modes

• Analysis Mode: Unweighted indicates that weights are not applied.
• If weights are needed, use Data Processing > Apply Weighted Analysis.

Best Practices

• Always start with a Codebook review before analysis.
• Tag key variables to organize large datasets.
• Use Notes to document analytical assumptions or data issues.
• Periodically export or back up analysis from the Analysis History.

Conclusion

The Chisquares Data Analysis module is a powerful tool for transforming survey responses into actionable insights. With an intuitive design and robust filtering options, users can efficiently analyze their data and make informed decisions.

For further assistance, access the Help? button in the lower-left for extended support.