VayuChat MCP

Load a CSV file into a pandas DataFrame.

Args: file_path: Path to the CSV file to load name: Optional name for the dataframe. If not provided, uses the filename without extension.

Returns: Summary of the loaded dataframe including shape, columns, and dtypes.

List all currently loaded dataframes with their basic info.

Returns: Summary of all loaded dataframes.

Get detailed information about a specific dataframe.

Args: name: Name of the dataframe to inspect

Returns: Detailed info including columns, dtypes, sample data, and statistics.

Execute Python code for data analysis.

The code has access to:

• All loaded dataframes by their names (e.g., 'air_quality', 'sales_data')

• pandas as 'pd'

• numpy as 'np'

• matplotlib.pyplot as 'plt'

For visualizations, use plt.savefig() or the code will automatically capture any open figures as base64 PNG images.

Args: code: Python code to execute

Returns: Output from the code execution, including any print statements and base64-encoded images for any generated plots.

Run a pandas query on a dataframe and return results.

This is a convenience method for simple queries. For complex analysis, use execute_code() instead.

Args: name: Name of the dataframe to query query: Pandas query string (e.g., "column > 100" or "city == 'Delhi'")

Returns: Filtered dataframe results.

Get statistical summary of a dataframe.

Args: name: Name of the dataframe columns: Optional list of specific columns to describe. If not provided, describes all numeric columns.

Returns: Statistical summary including count, mean, std, min, max, and quartiles.

Unload a dataframe from memory.

Args: name: Name of the dataframe to unload

Returns: Confirmation message.

Get a sample of rows from a dataframe.

Args: name: Name of the dataframe n: Number of rows to return (default: 10) random: If True, return random sample. If False, return first n rows.

Returns: Sample rows from the dataframe.

Compare weekday vs weekend values for a metric.

Args: name: Name of the dataframe value_column: Column containing the values to compare (e.g., 'PM2.5', 'sales') date_column: Column containing dates (default: 'date') day_of_week_column: Column containing day names (default: 'day_of_week') group_by: Optional column to group by (e.g., 'city', 'station')

Returns: Comparison of weekday vs weekend averages with statistics.

Compare a metric across different groups (e.g., cities, categories).

Args: name: Name of the dataframe value_column: Column containing values to compare group_column: Column containing groups (e.g., 'city', 'category') groups: Optional list of specific groups to compare. If None, uses all groups.

Returns: Statistical comparison across groups.

Analyze hourly patterns in the data.

Args: name: Name of the dataframe value_column: Column containing values to analyze hour_column: Column containing hour (0-23, default: 'hour') group_by: Optional column to group by (e.g., 'city')

Returns: Hourly pattern analysis with peak/off-peak hours.

Analyze correlations between numeric columns.

Args: name: Name of the dataframe columns: Optional list of columns to analyze. If None, uses all numeric columns. target: Optional target column to show correlations with (sorted by strength).

Returns: Correlation matrix or target correlations.

Analyze trends over time.

Args: name: Name of the dataframe value_column: Column containing values to analyze date_column: Column containing dates (default: 'date') period: Aggregation period - 'daily', 'weekly', 'monthly' (default: 'daily') group_by: Optional column to group by

Returns: Trend analysis with statistics.

Find top and bottom records by a value column.

Args: name: Name of the dataframe value_column: Column to rank by n: Number of top/bottom records (default: 5) group_by: Optional column to find top/bottom within each group

Returns: Top and bottom records.

Get values from a specific column.

Args: name: Name of the dataframe column: Column name unique: If True, return unique values. If False, return value counts. top_n: Limit to top N values (useful for columns with many unique values)

Returns: Column values or value counts.

Create a comparison chart across groups.

Args: name: Name of the dataframe value_column: Column containing values to plot group_column: Column containing groups (e.g., 'city') chart_type: Type of chart - 'bar', 'horizontal_bar', 'box' (default: 'bar') title: Optional chart title

Returns: Base64 encoded plot image.

Create a time series plot.

Args: name: Name of the dataframe value_column: Column containing values to plot date_column: Column containing dates (default: 'date') group_by: Optional column to create separate lines for each group title: Optional chart title

Returns: Base64 encoded plot image.

Create a distribution histogram.

Args: name: Name of the dataframe value_column: Column containing values to plot group_by: Optional column to create overlaid distributions bins: Number of histogram bins (default: 30) title: Optional chart title

Returns: Base64 encoded plot image.

Create an hourly pattern plot.

Args: name: Name of the dataframe value_column: Column containing values to plot hour_column: Column containing hour (0-23, default: 'hour') group_by: Optional column to create separate lines for each group title: Optional chart title

Returns: Base64 encoded plot image.

Create a weekday vs weekend comparison bar chart.

Args: name: Name of the dataframe value_column: Column containing values to compare day_of_week_column: Column containing day names (default: 'day_of_week') group_by: Optional column to group by (e.g., 'city') title: Optional chart title

Returns: Base64 encoded plot image.