📊 Correlation matrix · dataset tool
client‑side Pearson correlation · no upload · instant matrix · CSV paste
✅in‑browser
✅CSV paste
✅numeric only
✅no server
Correlation matrix (Pearson)
⚡ pure javascript · private · instant
🔢 first row = headers ✓
📐 numeric columns only auto
⚡ pearson r rounded 3
🧮 client‑side
📊 square matrix
ⓘ paste comma or tab-separated data — first row must be headers; all numeric columns are included automatically.
📋 Correlation matrix
0 × 0| — |
|---|
| ⬆️ paste data above to generate matrix |
Why client‑side matrix?
Sensitive data stays in your browser. Paste CSV with headers, we extract numeric columns and compute Pearson r. Perfect for exploratory analysis without uploading.
✓ Pearson r (listwise)
✓ handles missing
✓ copy as text
🔢📐
symmetric matrix
rounded to 3 decimals
What is a correlation matrix?
A correlation matrix is a table that shows the Pearson correlation coefficient between every pair
of numeric variables in a dataset. Each cell in the matrix contains a value between −1 and +1 that
describes how strongly and in what direction two variables are linearly related.
The matrix is always square and symmetric — the number of rows equals the number of columns (one for each
variable), and the value in row A, column B is always identical to the value in row B, column A. The diagonal
cells always show 1.000 because every variable is perfectly correlated with itself.
Correlation matrices are one of the first steps in exploratory data analysis (EDA). They help you quickly identify
which pairs of variables move together, which are independent, and which have a strong inverse relationship —
all in a single compact table.
💡 Example: In a dataset with columns for study hours, sleep hours, and exam score,
a correlation matrix would show you at a glance whether students who study more tend to score higher,
whether more sleep is linked to better scores, and whether studying more is associated with sleeping less.
What is the Pearson correlation coefficient (r)?
The Pearson r measures the strength and direction of the linear relationship between two continuous
numeric variables. It was developed by mathematician Karl Pearson in the late 19th century and remains the most
widely used measure of correlation in statistics, science, and data analysis.
The formula computes how much the two variables vary together (their covariance) relative to how much each varies
on its own (their standard deviations). The result is always a number between −1 and +1.
How to read the correlation values
| r value | Interpretation | Example |
|---|---|---|
| +1.000 | Perfect positive correlation | Variable always increases together |
| +0.7 to +0.99 | Strong positive correlation | Height and weight |
| +0.4 to +0.69 | Moderate positive correlation | Study hours and exam score |
| +0.1 to +0.39 | Weak positive correlation | Slight tendency to increase together |
| 0.000 | No linear correlation | Shoe size and exam score |
| −0.1 to −0.39 | Weak negative correlation | Slight tendency to move in opposite directions |
| −0.4 to −0.69 | Moderate negative correlation | Stress level and sleep quality |
| −0.7 to −0.99 | Strong negative correlation | Speed and journey time |
| −1.000 | Perfect negative correlation | Variables always move in exactly opposite directions |
How to use this correlation matrix tool
- Prepare your data — your dataset should have one row per observation and one column per variable. The first row should contain column headers (variable names). All data columns should contain numbers.
- Paste your CSV data — copy your data from Excel, Google Sheets, or any CSV file and paste it into the input box. The tool accepts comma-separated or tab-separated values. The first row is treated as headers automatically.
- Check column detection — the tool automatically identifies numeric columns and skips text columns (such as a row-index column with years or labels). Only numeric columns are included in the matrix.
- Read the matrix — each cell shows the Pearson r between the column variable and the row variable, rounded to 3 decimal places. The diagonal always shows 1.000. Look for cells close to +1 or −1 to find the strongest relationships.
- Copy or export — click Copy matrix (text) to copy the full matrix as plain text, or Export .csv to download it as a CSV file you can open in Excel or Google Sheets.
- Load sample data — click Load sample to see the tool in action with a built-in example dataset before entering your own data.
- Make sure the first row contains column names (e.g. "sales", "profit", "temperature"). The tool uses these as labels in the matrix.
- All values in numeric columns should be numbers only — no units, currency symbols, or text mixed in.
- A non-numeric first column (like a year or ID column) is automatically treated as a row index and excluded from the calculation.
- Missing values are handled using listwise deletion — any row with a missing value in either variable is excluded from that specific pair's calculation.
- For reliable results, aim for at least 10 to 20 rows of data. Correlation values from very small samples can be misleading.
Features
Instant calculation
The matrix is computed and displayed the moment you paste your data. No button press needed.
100% private
Your data never leaves your browser. All Pearson r calculations run in JavaScript with zero network calls.
CSV and tab paste
Paste directly from Excel, Google Sheets, or any CSV file. Comma and tab separators are both supported.
Pearson r (listwise)
Uses the standard Pearson correlation formula with listwise deletion for missing values.
Symmetric matrix
Output is a fully symmetric square matrix with 1.000 on the diagonal — standard academic format.
Auto column detection
Numeric columns are detected automatically. Text and index columns are excluded without any manual setup.
Copy and export
Copy the matrix as plain text or download it as a .csv file for use in reports, papers, or further analysis.
Works offline
Once the page loads, you can disconnect from the internet. The tool continues to work with no interruption.
Who is this tool for?
Correlation matrices are used across every field that works with numeric data. This tool is built for anyone
who needs fast, private, and reliable correlation analysis without installing software.
🏫 Students
Analyse survey data, lab results, or social science datasets for assignments, dissertations, and research projects.
🔬 Researchers
Run quick exploratory correlation checks on new datasets before moving to regression or multivariate analysis.
📉 Data analysts
Use as a fast sanity check on feature relationships before building machine learning models or dashboards.
📚 Teachers
Demonstrate Pearson correlation live in the classroom using real data. Students can run their own matrices on any device.
💼 Business analysts
Explore relationships between sales, costs, marketing spend, and other KPIs without sending data to a third-party tool.
💊 Economists
Check correlations between economic indicators such as GDP, inflation, unemployment, and interest rates.
Pearson r vs other correlation measures
There are several ways to measure correlation, each suited to different types of data. Here is how Pearson r
compares to the most common alternatives:
- Pearson r (this tool) — measures the strength of the linear relationship between two continuous numeric variables. Assumes both variables are approximately normally distributed and have no major outliers. Best for: scientific measurements, financial data, scores, and other continuous variables.
- Spearman rank correlation — a non-parametric alternative that measures monotonic relationships (not just linear ones). Works on ranked data and is more robust to outliers and non-normal distributions. Best for: ordinal data (e.g. survey ratings), or when Pearson assumptions are not met.
- Kendall tau — another non-parametric rank-based measure, less sensitive to errors in data and more suitable for very small samples. Less commonly used than Spearman in practice.
- Point-biserial correlation — used when one variable is continuous and the other is binary (0 or 1). For example, correlating exam score with pass/fail outcome.
📌 When to use Pearson r: Use Pearson when both variables are continuous numbers, your dataset
has at least 10 to 15 observations per variable, there are no extreme outliers, and you are interested in
linear relationships. If your data is ordinal or heavily skewed, consider Spearman instead.
How to interpret a correlation matrix
Reading a correlation matrix correctly is just as important as calculating it. Here is a step-by-step guide
to extracting meaningful insights from the output:
Step 1 — Identify the diagonal
The diagonal of the matrix always shows 1.000. This is expected and simply means each variable is
perfectly correlated with itself. You can ignore the diagonal when looking for relationships.
Step 2 — Focus on one triangle
Because the matrix is symmetric, the upper-right triangle and lower-left triangle contain the same values
in mirror image. You only need to read one side — typically the upper-right triangle above the diagonal.
Step 3 — Look for strong correlations
Start with cells showing values above +0.70 or below −0.70. These are strong relationships
worth investigating. Values near 0.00 suggest the two variables are not linearly related.
Step 4 — Watch for multicollinearity
In regression analysis and machine learning, very high correlations between predictor variables (above +0.90 or
below −0.90) can cause problems called multicollinearity. If two predictor variables are almost
perfectly correlated, consider removing one of them from your model.
Step 5 — Remember correlation is not causation
A high correlation coefficient tells you two variables move together — it does not tell you why.
A third variable (called a confounding variable or lurking variable) may be driving both.
Always combine correlation analysis with domain knowledge, theory, and further testing.
Common mistakes when using correlation matrices
- Including non-numeric columns — text, dates, and categorical columns should be removed or encoded before calculating correlations. This tool handles this automatically, but check that your data is clean before pasting.
- Using too few data points — a correlation of +0.90 from only 5 data points is not reliable. The more observations you have, the more trustworthy the r values are.
- Ignoring outliers — a single extreme data point can inflate or deflate Pearson r significantly. Always check your scatter plots alongside the correlation matrix.
- Treating all correlations as meaningful — with many variables, some correlations will appear significant purely by chance. Apply statistical significance testing (p-values) for formal analysis.
- Assuming linearity — Pearson r only captures linear relationships. Two variables can have a strong curved (non-linear) relationship and still show a Pearson r near 0. Use scatter plots to check.
- Confusing correlation with causation — the most common mistake. A high r value is a starting point for investigation, not a conclusion.
Frequently asked questions
What data format should I paste?
Paste comma-separated (CSV) or tab-separated data with the first row as column headers. You can copy directly from Excel or Google Sheets — tab-separated paste from spreadsheets works automatically. Each subsequent row should be one observation, and each column one variable.
What does a correlation of 0 mean?
A Pearson r of 0.000 means there is no linear relationship between the two variables. It does not mean the variables are completely unrelated — they could still have a strong non-linear (curved) relationship. Always visualise your data with a scatter plot alongside the matrix.
Why does the diagonal always show 1.000?
The diagonal represents the correlation of each variable with itself. Any variable is perfectly and positively correlated with itself, so the value is always exactly 1.000. You can ignore the diagonal when looking for meaningful relationships.
How are missing values handled?
This tool uses listwise deletion (also called complete case analysis). For each pair of variables, any row that has a missing value in either variable is excluded from that specific calculation. This is the standard approach for Pearson correlation and is the same method used by Excel's CORREL function.
How many variables can I include?
There is no hard limit. The tool generates a matrix for all numeric columns in your pasted data. For very large datasets with many columns, the matrix may become wide — use the CSV export and open it in Excel for easier reading of large matrices.
Is the result the same as Excel's CORREL function?
Yes. This tool uses the standard Pearson product-moment correlation formula, which is identical to Excel's CORREL function, Google Sheets' CORREL function, R's cor() function (with method = "pearson"), and Python's pandas DataFrame.corr() method (default settings).
Is my data safe to paste here?
Yes. All processing happens entirely inside your browser using JavaScript. No data is sent to any server, stored in a database, or shared with any third party. You can verify this by checking your browser's Network tab in developer tools — you will see zero data requests after the page loads. The tool also works offline once the page has loaded.
What is the difference between correlation and covariance?
Covariance measures how two variables change together but its value depends on the units of measurement, making it hard to interpret and compare. Pearson r is covariance divided by the product of both standard deviations, which standardises the result to a scale of −1 to +1 regardless of units. This makes correlation much easier to interpret and compare across different pairs of variables.
Why use a browser-based correlation matrix tool?
Correlation analysis often involves sensitive datasets — patient records, financial reports, student grades,
or proprietary business data. Uploading this data to an online service creates a privacy risk, even when
the service claims to be secure. A breach, a misconfigured server, or a change in privacy policy can expose
your data without warning.
This tool eliminates that risk entirely. All calculations happen locally in your browser using standard JavaScript.
There is no server, no upload, no account, and no cost. Your dataset is processed in the same environment as
the web page itself and is never transmitted anywhere. You remain in full control of your data at all times.
The tool is also instant. Because there is no round-trip to a server, correlation matrices for typical datasets
are computed in milliseconds. This makes it ideal for iterative data exploration — paste your data, check the
matrix, adjust your columns, and repeat — all without waiting for a page reload or API response.
Finally, the tool works offline. Once the page has loaded in your browser, you can disconnect from the internet
and continue working. This is useful in classrooms, research environments, or any situation where connectivity
is limited or restricted.