Data wrangling
Scrape data from the web, import it from standard formats, or pull it in via SQL with JDBC or ODBC. Match-merge, link, append, reshape, transpose, sort, filter. Stata handles Unicode, frames (multiple datasets in memory), BLOBs, regular expressions, and more, whether working with hundreds of thousands or even billions of data points.

Automated reporting and customizable tables
Use Markdown to create Word documents and HTML files with embedded Stata code, output, and graphs. Automate Word, PDF, or Excel reports with both high-level export capabilities and low-level fine-grained programmatic access to automate production of the documents your team needs. Customize tables to clearly communicate results, and export your tables to Word, PDF, HTML, LaTeX, Excel, or Markdown. Read more about Markdown, about Word Documents, about PDF documents, about Excel, or about customizable tables.

Visualization
Create graphs and customize them programmatically or interactively with the Graph Editor. Edits can even be recorded and "replayed" on other graphs for reproducibility. Export to industry standard formats suitable for web (SVG, PNG) or print (PDF, TIFF, EPS, PS).

Programming
Automate your entire workflow with both scripts and full-blown programming features like classes, structures, and pointers. A unique feature of Stata's programming environment is Mata, a fast and compiled matrix programming language. Of course, it has all the advanced matrix operations you need. It also has access to the power of LAPACK. What's more, it has built-in solvers and optimizers to make implementing your own estimator easier. And you can leverage all of Stata's estimation features and other features from within Mata.

PyStata—Python integration
Interact Stata code with Python code. You can seamlessly pass data and results between Stata and Python. You can use Stata within Jupyter Notebook and other IPython environments. You can call Python libraries such as NumPy, matplotlib, Scrapy, scikit-learn, and more from Stata. You can use Stata analyses from within Python.

Interoperability
Connect to external code via Python, Java, and C++ plugins. Write Python or Java code directly within your Stata code. Control Stata via Jupyter Notebook, OLE Automation, or call it in batch mode. Write custom SQL statements with JDBC and ODBC to extract from or populate databases. Access H2O clusters. Read more about PyStata—Python integration, Jupyter Notebook, Java integration, Java plugins, C/C++ plugins, H2O integration, and OLE Automation.

Statistics and modeling
Incorporate state-of-the-art statistical models and results in your workflow. Find groups in your data using unsupervised techniques including cluster analysis, principal components, factor analysis, multidimensional scaling, and correspondence analysis. Understand your groups even better using latent class analysis. When your analysis calls for supervised techniques, Stata has flexible nonparametric methods and an array of regression models from linear and logistic models to mixture models. Stata keeps up when your data call for special techniques. You have access to methods that understand and take advantage of the structure in time series, panel data, survival data, complex survey data, spatial data, and multilevel data. Stata provides the most approachable implementations of Bayesian methods and structural equation modeling available anywhere. You can request bootstrap methods for virtually any estimator. When your analysis calls for it, Stata automates other replication methods and simulations.

Reproducibility
Stata is the only software for data science and statistical analysis featuring a comprehensive version control system that ensures your code continues to run, unaltered, even after updates or new versions are released. No need to keep around multiple legacy installations to avoid breaking your system; Stata code from 25 years ago can still be run without modification. Datasets, graphs, scripts, programs, and more are 100% cross-platform and backward compatible.

Lasso
Use lasso and elastic net for model selection and prediction. And when you want to estimate effects and test coefficients for a few variables of interest, inferential methods provide estimates for these variables while using lassos to select from among a potentially large number of control variables. You can even account for endogenous covariates. Whether your goal is model selection, prediction, or inference, you can use Stata's lasso features with your continuous, binary, count, or time-to-event outcomes.