Skip to main content
Glama
ChenJellay

Data Analytics MCP Toolkit

by ChenJellay
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Hybrid

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
PYTHONPATHNoEnvironment variable to ensure the 'src' directory is in the Python search path for module resolution.

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": false
}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}
experimental
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
load_dataA
Ingest data from CSV or JSON (inline string or URL). Returns data_id and schema summary.
Use the returned data_id in subsequent tools (clean_data, plot_*, train_*, etc.).
clean_dataB
Clean dataset: optionally drop NA rows and z-score normalize numeric columns.
Updates the dataset in place; returns data_id and row count.
plot_barC

Bar chart: x_column as categories, y_column as values (or count of x if y_column omitted).

plot_lineC

Line chart: x_column on x-axis, one or more y_columns as lines.

plot_scatterC

Scatter plot of x_column vs y_column.

plot_histogramC

Histogram of a numeric column (distribution).

plot_boxC

Box plot: single numeric column, or all numeric columns if column is omitted.

plot_heatmapC

Heatmap of correlation matrix. If columns omitted, uses all numeric columns.

train_test_splitB
Split dataset into train and test. Returns train_data_id and test_data_id
for use in train_* and evaluate_* tools.
train_linear_regressionC

Fit a linear regression model. Returns model_id for evaluate_regression.

train_logistic_regressionB

Fit a logistic regression classifier. Returns model_id for evaluate_classification.

train_kmeansC

Fit K-means clustering. Returns model_id for evaluate_clustering.

evaluate_regressionC

Compute MSE and R² for a regression model on test data.

evaluate_classificationC

Compute accuracy for a classification model on test data.

evaluate_clusteringC

Compute silhouette score for a clustering model on test data.

run_analyticsB
High-level tool: describe what you want (e.g. "show distribution of sales",
"predict price from square_feet", "cluster into 4 groups") and provide the data
(CSV/JSON string or URL). The server picks the right pipeline and returns
either a chart (chart_base64, chart_type) or ML metrics and model summary.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription
list_pipelinesList available analytics pipelines with short descriptions.
pipeline_visualizationSteps: 1) load_data(source, format) 2) clean_data(data_id) 3) plot_histogram/plot_bar/plot_line/plot_scatter/plot_box/plot_heatmap(data_id, column(s)). Or use run_analytics(intent, data_source).
pipeline_regressionSteps: 1) load_data 2) clean_data 3) train_test_split(data_id, target_column) 4) train_linear_regression(train_data_id, target_column) 5) evaluate_regression(model_id, test_data_id). Or use run_analytics(intent, data_source) with intent like 'predict Y from X'.
pipeline_classificationSteps: 1) load_data 2) clean_data 3) train_test_split 4) train_logistic_regression 5) evaluate_classification. Or use run_analytics with intent like 'classify' or 'predict category'.
pipeline_clusteringSteps: 1) load_data 2) clean_data 3) train_kmeans(data_id, n_clusters) 4) evaluate_clustering(model_id, data_id). Or use run_analytics with intent like 'cluster into k groups'.
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

B3.3/5.0

Scored across 16 tools

Disambiguation4/5

Most tools have distinct purposes, with clear separation between data loading/cleaning, plotting, model training, and evaluation functions. However, 'run_analytics' overlaps significantly with the specialized tools, as it can perform many of the same functions through a single interface, which could cause confusion about when to use it versus the specific tools.

Naming Consistency5/5

Tool names follow a consistent verb_noun pattern throughout, with clear and predictable naming conventions. All tools use snake_case, and verbs like 'clean', 'evaluate', 'load', 'plot', 'run', 'train', and 'train_test_split' are applied consistently to their respective nouns, making the set highly readable and predictable.

Tool Count4/5

With 16 tools, the count is slightly high but reasonable for a comprehensive data analytics toolkit. It covers a broad range of functions from data ingestion to visualization and machine learning, though it might be borderline heavy for some use cases. Each tool appears to earn its place without obvious redundancy, except for the overlap with 'run_analytics'.

Completeness5/5

The tool set provides complete coverage for a data analytics workflow, including data loading, cleaning, splitting, multiple types of plots, training for classification, regression, and clustering models, and corresponding evaluation metrics. The inclusion of 'run_analytics' as a high-level tool further ensures no gaps, offering a flexible alternative for common tasks.

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources