Alteryx for Data Scientists
Get to your next breakthrough faster. Accelerate machine learning modeling and focus on the work that interests you most with Analytic Process Automation (APA).
Combine custom code with automation
Automate the repetitive, customize the creative, and deeply explore your data. Whether via your own code or configurable analytic building blocks, you can record and replay any data preprocessing, exploration, and feature engineering steps on any data set, on demand.
Stand up a working model in minutes. See variable relationships and distributions instantly. Select and compare algorithm performance with just a few clicks. Train a model and capture an end-to-end machine learning pipeline in code and visual documentation.
Operationalize faster for
Experience self-service model deployment with no recoding required. Deploy in a secure cloud environment, on-premises behind your own firewall, or in a hosted environment. Easily integrate models into existing applications via standard REST API requests.
- Data Access, Quality, and Cleansing
- Feature Engineering and Modeling Approaches
- Analysis Visualization and Documentation
- Real-time Model Deployment
- Model Management and Monitoring
Data Access, Quality, and Cleansing
Connecting to data sources is simple, whether your data is on-premises, in the cloud, Big Data, small, structured, semi-structured, or unstructured. In-database analytic building blocks leverage the power of the cloud to run analytics against your biggest data sources. Plus, Analytic Process Automation makes it simple to enrich your existing sources with robust third-party data for deeper location and business insights.
Analytic workflows let you pinpoint any step to take a new prep or cleanse action for optimal flexibility. Plus, you can easily replicate actions on different data sets with the built-in repeatability of analytic workflows.
Feature Engineering and Modeling Approaches
Explore your data while you create, assess, and select features with Analytic Process Automation — a visual programming interface that offers independent transformations in analytic building blocks to power precise, granular changes. Leverage quick configurations in prebuilt building blocks alongside your own R or Python code in the same analytic workflow.
Quickly prototype machine-learning models and pipelines with automated model-training building blocks for the most popular supervised and unsupervised predictive, prescriptive, and time-series models. Experience feature engineering and model training with the unprecedented ease and speed of analytic workflows — which can leverage both your own code and Alteryx analytic building blocks for the right blend of custom and out-of-the-box transformations.
Analysis Visualization and Documentation
Analytic Process Automation helps you easily visualize your data throughout your entire problem-solving and modeling journey with a data profile available at every step.
Explore your data in an onboard results grid at any point in your analytic workflow to make faster decisions about your next step. Automatically generate charts, tables, and reports from any step in your process — even dashboards that non-technical stakeholders can drill into.
Transparency becomes automatic with analytic workflows that visually display every step in your analysis with customizable annotations — whether you’re using your own code, prebuilt analytic building blocks, or any combination thereof. Plus, dynamic batch reporting is available at a click to automatically create and distribute multiple easily configurable reports from any analysis.
Real-time Model Deployment
Deploy R, Python, or even low-code models quickly and reliably with a self-service Analytic Process Automation system that cuts out the delays and risks of recoding. Deploy where you need to, whether in a secure cloud environment, on-premises behind your own firewall, or in a hosted environment.
Models are easily embedded in business- and consumer-facing applications via standard REST API requests — code snippets are even provided for you. Power real-time predictions in low-latency applications with included high availability and failover. Run on a clustered architecture that can easily scale out for running larger workloads. Effortlessly replicate models to accommodate seasonal or permanent increases in throughput.
Model Management and Monitoring
Ensure models deliver the best performance for ongoing effectiveness in production with a live view of model status, model performance, and system health. Administer your models and servers from a centralized Analytic Process Automation system that offers insight into model uptime, resource consumption, and responses. Plus, organized collaboration within and across data science teams becomes easy with built-in model version control and tracking.
With Analytic Process Automation, the model production, management, and deployment process has never been smoother. Say goodbye to long, expensive custom model deployments, loss of control over production models, and total reliance on IT.