R & Python Research Lab for publication-ready statistics and graphs
Run real R 4.6 and Python 3.14 right in your browser. Upload your data, pick a ready-made analysis (t-test, ANOVA, regression, survival, ROC, PCA, clustering, machine learning and genomics: volcano, Manhattan, enrichment, ORF finder) and download journal-quality figures at 300 or 600 dpi.
Tip: drag and drop a file here. Click a column name to insert it into your code. The first row must contain column names.
Why researchers use this lab
The same R and Python used in RStudio and Jupyter, compiled to run in your browser with WebAssembly. Install CRAN packages with need("package").
Clean publication themes, colour-blind friendly palette and export as 300/600 dpi PNG or TIFF, or vector SVG and PDF, at the exact size a journal asks for.
Your data is processed on your own computer. Nothing is uploaded to our server, so unpublished and patient data stay confidential.
Every analysis shows the exact code. Save the script, add it to your supplementary material and rerun it any time in RStudio or Jupyter.
How to use it in 4 steps
Upload a CSV or Excel file, paste from Excel, or start with our sample dataset.
Choose from 167 templates in R or Python: tests, models, survival, ROC, PCA, genomics and 49 machine-learning methods.
Change the column names at the top of the code (click a column name to insert it).
Press Run, read the statistics, then download figures in the format your journal needs.
Cite the language and packages you used, not this website. Typical wording for the Methods section:
Statistical analyses were performed in R version 4.6.0 (R Core Team, 2026) using the ggplot2 and survival packages, run through webR. / Analyses were performed in Python 3.14 with SciPy, statsmodels and matplotlib (via Pyodide).Run citation() or citation("ggplot2") in the R tab to get the exact references.
Frequently asked questions
Is this really R and Python, like RStudio and Jupyter?
Yes. The R tab runs the official R interpreter (version 4.6) through webR, and the Python tab runs CPython 3.14 through Pyodide. Both are compiled to WebAssembly, so they run inside your browser instead of on a server. The code you write here also works unchanged in RStudio and Jupyter.
Which R packages and Python libraries can I use?
In R you can install most popular CRAN packages with need("package"), including ggplot2, dplyr, tidyr, survival and many more. In Python, NumPy, pandas, SciPy, matplotlib, statsmodels and scikit-learn are available. A few packages that need special system libraries may not work in the browser.
How do I get a 300 dpi or 600 dpi TIFF for my journal?
After running an analysis, choose TIFF 300 dpi or TIFF 600 dpi in the Download figures bar, set the width and height in inches (for example 3.5 in for a single column or 7 in for a double column) and click Download.
Can I analyse genomics data here?
Yes. The genomics templates cover RNA-seq (differential expression from counts, volcano and MA plots, heatmaps, sample PCA, QC), GWAS (Manhattan and QQ plots with genomic inflation), gene sets (enrichment dot plots, UpSet plots, over-representation tests) and DNA sequences (FASTA statistics, GC content, six-frame ORF finder). Each template loads a matching sample dataset; upload your own file in the same format (for example a DESeq2 results table, a count matrix or a FASTA file). For final RNA-seq statistics in a paper, we recommend confirming results with DESeq2, edgeR or limma in RStudio.
Which machine-learning methods are included?
49 templates for engineering, life science and biotechnology. Classification: logistic regression, k-NN, SVM, naive Bayes, decision trees, random forest, gradient boosting, XGBoost, LightGBM and neural networks, with ROC, precision-recall, calibration, multi-class ROC, hyperparameter tuning, learning curves and nested cross-validation. Regression: LASSO, elastic net, polynomial, SVR, Gaussian process, ANN and boosting with partial dependence and prediction intervals (conformal and quantile). Advanced: stacking and voting ensembles, imbalanced data (SMOTE, class weights, threshold tuning), feature selection (RFECV, mutual information, Boruta-style), biomarker stability selection for omics, Bayesian optimisation to suggest next experiments, explainable AI (SHAP waterfall, dependence, ICE), causal treatment-effect estimation (T- and X-learner), semi-supervised learning, autoencoder and isolation-forest anomaly detection, six clustering algorithms, manifold learning (PCA, kernel PCA, Isomap, LLE, MDS, t-SNE), random survival forests, chemometrics (PLS, PLS-DA, SNV/MSC/Savitzky-Golay), response surface methodology, ARIMA and machine-learning forecasting, and DNA k-mer sequence classification. Each loads a matching sample dataset so you can try it immediately.
Can I run deep learning (TensorFlow or PyTorch)?
Not yet: TensorFlow and PyTorch are not available inside the browser engine. The lab includes multilayer neural networks and autoencoders through scikit-learn, which suit typical tabular research datasets. For image or very large deep-learning models, use the Python code from the lab as a starting point in Google Colab or Jupyter.
Is my data safe?
Your data stays on your computer. The lab only downloads the R or Python engine and packages; your file is read and analysed locally in your browser and is never sent to dailyfirst.in.
Why is the first run slow?
The first time, your browser downloads the R or Python engine (about 30 to 40 MB) and any packages a template needs. Your browser caches them, so later runs start in seconds.
Does it work on mobile?
It works in up-to-date Chrome, Safari, Edge and Firefox, including on recent phones and tablets, but a laptop or desktop is best for larger datasets.