Correlation and Relationship

Endpoint: POST /correlation-relationship

Request Parameters

File Upload

• file (required): The file to be processed. Supported formats include CSV, Excel, etc.

Correlation Method

• correlation_method (optional): The method used to compute correlations between numerical columns. Default is pearson. Other options include:
  • pearson: Pearson correlation coefficient (default).
  • spearman: Spearman rank-order correlation.

Numerical Columns

• numerical_columns (optional): Comma-separated list of numerical columns to include in the correlation analysis. If omitted, all numerical columns are used.

Cross-Tabulation Columns

• cross_tab_columns (optional): Comma-separated list of two categorical columns to compute a cross-tabulation. If omitted or if more than two columns are provided, no cross-tabulation is performed.

Include Graphs

• include_graphs (optional): Boolean to specify if graphs should be included in the response. Default is false.

Graph Types

• graph_types (optional): Comma-separated list of graph types to include in the analysis. Default is []. Available options:
  • heatmap: Displays a heatmap of the correlation matrix.
  • stacked_bar: Displays a stacked bar chart for the cross-tabulation data.
  • clustered_bar: Displays a clustered bar chart for the cross-tabulation data.

Example Request

curl -X POST "http://localhost:8000/correlation-relationship" \
-F "file=@data.csv" \
-F "correlation_method=spearman" \
-F "numerical_columns=Age,Income" \
-F "cross_tab_columns=Gender,Education" \
-F "include_graphs=true" \
-F "graph_types=heatmap,stacked_bar"

Analysis Components

1. Correlation Analysis: Computes the correlation matrix for numerical columns using the specified correlation method (pearson or spearman).
2. Cross-Tabulation: Computes a cross-tabulation (contingency table) for two categorical columns if provided. This is useful for understanding the relationship between categorical variables.
3. Graphs: Optionally generates visualizations based on the correlation matrix and cross-tabulation:
   • Heatmap: Visualizes the correlation matrix as a heatmap.
   • Stacked Bar: Visualizes the cross-tabulation data as a stacked bar chart.
   • Clustered Bar: Visualizes the cross-tabulation data as a clustered bar chart.

Example Response

{
  "correlation_matrix": {
    "Age": {"Income": 0.45, "Outcome": 0.32},
    "Income": {"Age": 0.45, "Outcome": 0.60},
    "Outcome": {"Age": 0.32, "Income": 0.60}
  },
  "cross_tabulation": {
    "Gender": {"Male": {"Bachelor": 50, "Master": 30}, "Female": {"Bachelor": 40, "Master": 20}}
  },
  "graphs": {
    "heatmap": [
      [1, 0.45, 0.32],
      [0.45, 1, 0.60],
      [0.32, 0.60, 1]
    ],
    "stacked_bar": {
      "Male": [50, 30],
      "Female": [40, 20]
    },
    "clustered_bar": {
      "Male": [50, 30],
      "Female": [40, 20]
    }
  }
}