Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/davi-huanuco/python-matriz-correlacion/llms.txt

Use this file to discover all available pages before exploring further.

The correlation matrix uses the Pearson correlation coefficient to measure the linear relationship between any two indicators. The result is a value in the range −1 to 1, where 1 means a perfect positive relationship, −1 means a perfect inverse relationship, and 0 means no linear relationship. The function that computes this is calcular_correlacion in data_store.py.

​
Function signature

def calcular_correlacion(datos: dict, codigo_a: str, codigo_b: str) -> float | None
​
datos
dict
required
The full application data dictionary, as returned by cargar_datos. Must contain "meses" and "indicadores".
​
codigo_a
str
required
The indicator code for the first series (row in the matrix).
​
codigo_b
str
required
The indicator code for the second series (column in the matrix).
Returns a float in the range −1 to 1, or None when the correlation cannot be computed (see edge cases below). The UI displays None as "-".

​
Algorithm

The function follows these steps:
1

Collect shared valid pairs

For each month in datos["meses"], the function checks whether both indicators have a numeric value (not null). Only months where both values are integers or floats are included. This list of (value_a, value_b) tuples is called pares.
2

Require at least two pairs

If pares has fewer than 2 elements, the function returns None. You need at least two data points to compute a meaningful correlation.
3

Compute the means

The arithmetic mean is computed separately for each series:
  • promedio_a = sum(lista_a) / len(lista_a)
  • promedio_b = sum(lista_b) / len(lista_b)
4

Compute the numerator

The numerator is the sum of the products of deviations from the mean:numerador = Σ (a − promedio_a) × (b − promedio_b)
5

Compute the denominator

The denominator is the square root of the product of the total squared deviations for each series:denominador = √( Σ(a − promedio_a)² × Σ(b − promedio_b)² )
6

Guard against zero denominator

If denominador equals 0 — which occurs when one or both series have no variance (all values are identical) — the function returns None to avoid a division-by-zero error.
7

Return the coefficient

Returns numerador / denominador, a float in the range −1 to 1.

​
Source code

def calcular_correlacion(datos, codigo_a, codigo_b):
    valores_a = datos["indicadores"][codigo_a]["valores"]
    valores_b = datos["indicadores"][codigo_b]["valores"]
    pares = []

    for mes in datos["meses"]:
        valor_a = valores_a.get(mes)
        valor_b = valores_b.get(mes)
        if isinstance(valor_a, (int, float)) and isinstance(valor_b, (int, float)):
            pares.append((valor_a, valor_b))

    if len(pares) < 2:
        return None

    lista_a = [par[0] for par in pares]
    lista_b = [par[1] for par in pares]
    promedio_a = sum(lista_a) / len(lista_a)
    promedio_b = sum(lista_b) / len(lista_b)
    numerador = sum(
        (valor_a - promedio_a) * (valor_b - promedio_b)
        for valor_a, valor_b in pares
    )
    suma_a = sum((valor_a - promedio_a) ** 2 for valor_a in lista_a)
    suma_b = sum((valor_b - promedio_b) ** 2 for valor_b in lista_b)
    denominador = math.sqrt(suma_a * suma_b)

    if denominador == 0:
        return None
    return numerador / denominador

​
When None is returned

The function returns None in two cases:
  • Fewer than 2 shared months with data — there is not enough overlap between the two indicators to compute a statistically meaningful result.
  • Zero denominator — at least one of the indicators has the same value in every recorded month, giving it zero variance. A correlation with a constant series is undefined.
In views.py, the crear_tabla_correlacion function renders None as "-" in the matrix cell: 
correlacion = calcular_correlacion(datos, codigo_fila, codigo_columna)
texto = "-" if correlacion is None else f"{correlacion:.2f}"

​
Interpreting the result

RangeInterpretation
0.90 to 1.00Very strong positive correlation
0.70 to 0.89Strong positive correlation
0.40 to 0.69Moderate positive correlation
0.00 to 0.39Weak or no positive correlation
−0.39 to 0.00Weak or no negative correlation
−0.69 to −0.40Moderate negative correlation
−0.89 to −0.70Strong negative correlation
−1.00 to −0.90Very strong negative correlation
The diagonal of the matrix (an indicator correlated with itself) always returns 1.00 because the shared pairs are identical and the deviations move in perfect lockstep.

Build docs developers (and LLMs) love

Get started for freeTalk to us