Statistical analysis
without the setup
Tell your AI assistant what test you need. MCP Analytics runs a statistical test that fits your question on your data and returns a report with the method named, charts, p-values and a plain-English reading. No R install and no SPSS license to set up.
machine-learning models
and describe the test
with its code shown
for papers & presentations
See what you get before you upload
Two reports built from a student-performance sample dataset. Each shows the method, the charts and the written reading, and you can open the PDF download inside the report.
Student scores compared across groups (ANOVA)
Descriptive statistics by group, score distributions and the test result for differences between groups.
Exam score regression
Exam scores modeled from student characteristics, with coefficients, fit statistics and diagnostics.
Parametric & Non-Parametric Tests
The core tests you'll need for any research methods course, thesis, or dissertation. Each report includes test statistic, p-value, effect size, and written interpretation.
Independent Samples t-test
Compare means between two groups. Returns t-statistic, degrees of freedom, p-value, Cohen's d, and confidence interval with plain-English interpretation.
One-Way ANOVA
Test for differences across three or more groups. Includes F-statistic, post-hoc Tukey comparisons, and effect size (η²) with assumption checks.
ANCOVA
Analysis of covariance: compare group means while controlling for a continuous covariate. Adjusts group means for a measured covariate such as a baseline score; it does not remove unmeasured confounders.
Chi-Square Test
Test independence between two categorical variables. Returns χ² statistic, degrees of freedom, p-value, Cramér's V effect size, and contingency table visualization.
Mann-Whitney U Test
Non-parametric alternative to the t-test when normality can't be assumed. Rank-based comparison with effect size r and bootstrapped confidence intervals.
Kruskal-Wallis Test
Non-parametric ANOVA alternative. Compares medians across 3+ groups with Dunn's post-hoc testing and effect size η²H.
Predictive Modeling
From simple linear regression to regularized models to classification: fully interpreted outputs with diagnostic plots, coefficients, and model fit statistics.
Linear Regression
Ordinary least squares with coefficient table, R², F-statistic, residual diagnostics, and multicollinearity checks (VIF). Full model summary.
Lasso Regression
Shrinks weak predictors to zero so the model keeps only the useful ones (L1 regularization). λ chosen by cross-validation, coefficient shrinkage plot, and sparse model interpretation.
Ridge Regression
Handles predictors that are strongly correlated with each other (L2 regularization). Shrinks coefficients toward zero without eliminating them, with optimal λ via cross-validation.
Logistic Regression
Binary or multiclass classification with odds ratios, ROC curve, AUC, confusion matrix, and probability calibration plots.
Naïve Bayes
Probabilistic classifier with feature independence assumption. Posterior probabilities, feature importance, and classification accuracy report.
Random Forest
Ensemble decision trees with variable importance, partial dependence plots, out-of-bag error estimates (error measured on rows each tree did not see), and tuning parameter selection.
Dimensionality Reduction & Unsupervised Learning
Explore structure in complex datasets. PCA, clustering, and correlation analysis for survey data, experimental results, and exploratory research.
Principal Component Analysis
PCA with scree plot, component loadings, biplot, and cumulative variance explained. Determine how many components to retain using the Kaiser criterion (keep components with eigenvalue above 1) and parallel analysis.
K-Means Clustering
Partition data into k groups with silhouette analysis for optimal k selection. Cluster profiles, within-cluster SS, and visualization plots.
DBSCAN Clustering
Density-based clustering that handles arbitrary shapes and noise. Labels outliers as noise instead of forcing them into a cluster, no need to pre-specify k.
Correlation Analysis
Full correlation matrix with Pearson, Spearman, or Kendall coefficients. Heatmap visualization, significance stars, and pairwise scatter plots.
Holm-Bonferroni Correction
Multiple comparisons correction to control family-wise error rate. Input your raw p-values and get adjusted values with interpretation of which results survive correction.
Intraclass Correlation (ICC)
Inter-rater reliability with ICC(2,1) and ICC(2,k) estimates. Agreement vs consistency distinction, confidence intervals, and ICC estimates with confidence intervals; see Koo and Li (2016) for how to choose and report an ICC.
Temporal Analysis
For longitudinal data, panel studies, and anything with a date column. Forecast future values or model time-to-event outcomes in survival studies.
ARIMA Forecasting
Auto-ARIMA with AIC-based order selection, ACF/PACF plots, residual diagnostics, and forecast with confidence bands.
Prophet Decomposition
Trend + seasonality + holiday decomposition using Prophet (Taylor and Letham, 2018). Interpretable components, changepoint detection, and uncertainty intervals.
Cox Proportional Hazards
Survival regression modeling time-to-event outcomes with covariates. Hazard ratios, Kaplan-Meier curves, log-rank tests, and proportional hazards assumption diagnostics.
Trend Analysis
Moving averages, Mann-Kendall trend test, Sen's slope estimator, and trend line visualization. Detect significant monotonic trends in panel data.
XGBoost
Gradient boosting with hyperparameter tuning, SHAP feature importance, learning curves, and calibrated probability outputs.
Data Profiler
Automated EDA on any CSV: distributions, missing values, outliers, correlations, and data type recommendations. Start every analysis here.
Same statistics. Zero setup.
Every model runs in R, with its code shown.
- Runs in your AI assistant
- No install: upload CSV, describe the test
- R-backed, with the code shown
- AI-written interpretation included in every report
- PDF download of each report for papers & presentations
What it can be used for
From intro statistics homework to dissertation analysis. These are examples of the kinds of questions it fits, not testimonials.
Thesis / Dissertation Data Analysis
You've collected survey responses or experimental data. You need proper tests with effect sizes and assumption checks, with the method named.
Experimental Methods Coursework
Weekly problem sets requiring specific tests. Upload your data file, say what test you need, get the output with full workings, cite the method in your write-up.
Machine Learning Projects
Need to compare classifiers, check a model on held-out data, or demonstrate feature importance for a project? Get outputs you can cite and re-run without writing a line of sklearn.
Economics & Finance Research
Time series, panel data, regression with controls, economics datasets need rigorous modeling. Get regression tables with the method named.
Run your first test on your own CSV
Upload your CSV, describe what you need, get a complete statistical report.
Where to start
One email with your data and your question. We take it from there.
MCP Analytics
Data science by email
Send us your data and your question, and we will reply.
Try it on your own data
30 days free.
Answers, analytics and custom analytics. No card, no call, nothing to install.
30 days free. →