Calculators
Bray–Curtis Dissimilarity Calculator
Calculate Bray–Curtis dissimilarity and similarity between non-negative numeric vectors privately in your browser.
Enter equal-length numeric vectors, separated by spaces, commas or semicolons. Use non-negative values; scientific notation is supported. Each vector allows 10000 numbers and 200000 characters. Paired zeros contribute zero; two all-zero vectors are undefined. Calculation runs locally in your browser.
How Bray–Curtis dissimilarity works
Bray–Curtis dissimilarity is Σ|aᵢ − bᵢ| / Σ(aᵢ + bᵢ) for non-negative values. It ranges from 0 for identical vectors to 1 when no coordinate is positive in both vectors. For A = (1, 2, 0) and B = (2, 4, 0), the result is 3/9 = 1/3. Similarity is 1 minus dissimilarity. Coordinates must represent the same categories in the same order, with consistent units. Paired zeros contribute nothing to either sum; two all-zero vectors are undefined and rejected. Unlike Canberra distance, this calculation divides summed absolute differences by total abundance rather than summing per-coordinate ratios. Multiplying both vectors by the same positive constant preserves the result; scaling each vector independently can change it. This dissimilarity does not generally satisfy the triangle inequality. Enter 1 to 10000 finite non-negative numbers per vector, separated by spaces, commas or semicolons. Scientific notation is allowed; commas are delimiters, not thousands separators. Each input allows 200000 characters. Values that underflow to zero are rejected. Shared scaling reduces overflow; negligible contributions can round away in browser floating-point arithmetic. Results show up to 12 significant digits. All processing stays in your browser.
Frequently Asked Questions
Everything you need to know about this tool, how it works, and privacy.
What is the Bray–Curtis formula?
Divide summed absolute coordinate differences by the sum of all values in both vectors. For (1, 2, 0) and (2, 4, 0), the result is 1/3.
What do zero and one mean?
Zero means identical vectors; one means no shared positive coordinates. Similarity equals 1 minus dissimilarity.
Are negatives and zeros allowed?
Values must be non-negative. Paired zeros contribute nothing. Two all-zero vectors are undefined; one zero vector paired with a nonzero vector gives dissimilarity 1.
How does this differ from Canberra distance?
Canberra sums coordinate-level relative differences. Bray–Curtis uses one ratio of summed differences to total abundance, giving larger values more influence.
Are vectors normalized automatically?
No. Use consistent units and matching coordinate order. Scaling both vectors by the same positive factor preserves the result; normalizing each independently may change it.