Top 8 predictive analytics tools compared

Do you want to know what the future may bring? Predictive analysis tools have an answer. Are they right? Sometimes. But sometimes can often be more than enough if the prediction can help your enterprise plan better, spend more wisely, and deliver more prescient service for your customers.

Predictive analytics tools blend artificial intelligence and business reporting. The tools include sophisticated pipelines for gathering data from across the enterprise, add layers of statistical analysis and machine learning to make projections about the future, and distill these insights into useful summaries so that business users can act on them.

The quality of predictions depends primarily on the data that goes into the system — the old slogan from the mainframe years, “garbage in, garbage out”, still holds today. But there are deeper challenges because predictive analytics software can’t magically anticipate moments when the world shifts gears and the future bears little relationship to the past. Still, the tools, which operate largely by ascertaining patterns, are growing increasingly sophisticated.

Working with dedicated predictive analytics tools is often relatively easy, at least compared to programming your own tools from scratch. Most tools offer visual programming interfaces that enable users to drag and drop various icons optimized for data analysis. It helps to understand coding and to think like a programmer, but the tools do make it possible to generate sophisticated predictions with just a few mouse clicks. If you need more, adding a bit of custom code can usually solve many common issues.

Alteryx Analytics Process Automation

The goal of Alteryx’s Analytic Process Automation (APA) platform is to help you build a pipeline that cleans data before applying the best data science and machine learning algorithms. A high level of automation encourages deploying these models into production to generate a constant stream of insights and predictions. The visual IDE offers more than 300 options that can be joined together to form a complex pipeline. One of the strengths of APA is its collection of deep integration with other data sources, such as geospatial databases or demographic data, to enrich the quality of your own data set.

Highlights:

• A very good solution for data scientists who must automate a complex collection of data sources to produce multiple deliverables
• For deployment locally or in the Alteryx cloud
• Includes many robotic process automation (RPA) tools for handling chores such as text recognition or image processing
• Designed to drive insights to multiple customers who might want data presented as dashboards, spreadsheets, or some other custom platform
• Pricing for tools such as the Designer starts at $5,195 per user, per year. Extras are priced by the sales team. Free trials and open source options are available.

AWS SageMaker

Amazon’s main AI platform is well-integrated with the rest of the AWS fleet so you can analyze data from one of cloud vendor’s major data sources and then deploy it to run either in its own instance or as part of a serverless lambda function. SageMaker is a full-service platform with data preparation tools such as the Data Wrangler, a nice presentation layer built out of Jupyter notebooks, and an automated option called Autopilot. Visualization tools can help users understand what’s going on with just one glance.

Highlights:

• Full integration with many parts of the AWS ecosystem makes this a great option for AWS-based operations
• Serverless options for deployment allow costs to scale with usage
• A marketplace facilitates buying and selling models and algorithms with other SageMaker users
• Integration with various AWS databases, data lakes, and other data storage options make working with big datasets simple
• Pricing is generally tied to the size of the computing resources used to support your calculations. A generous free tier makes it possible to experiment.

H2O.ai AI Cloud

Turning good artificial intelligence algorithms into productive insights is the main goal of H2O.ai’s AI Cloud. Its “Driverless AI” offers an automated pipeline for ingesting data and studying the most salient features. A collection of open source and proprietary feature engineering tools help focus the algorithms on the most important parts of the data. The results appear in a collection of dashboards or automated graphical visualizations.

Highlights:

• Focus on AI is best for problems that require complex solutions that adapt to incoming data.
• Tools range from AI Cloud for creating large, data-driven pipelines to open source, Python-based Wave that helps desktop users create real-time dashboards
• Runs natively on premises or in any cloud
• Core platform is fully open source
• Pricing for enterprise support and cloud options available from the sales team

IBM SPSS

Statisticians have been using IBM’s SPSS to crunch numbers for decades. The latest version includes options for integrating newer approaches such as machine learning, text analysis, or other AI algorithms. The Statistics package focuses on numerical explanations of what happened. SPSS Modeler is a drag-and-drop tool for creating data pipelines that lead to actionable insights.

Highlights:

• Ideal for large, traditional organizations with big data flows
• Integrated with other IBM tools such as Watson Studio
• Leverages larger initiatives such as IBM’s push for Trustworthy AI
• Pricing begins at $499 per user, per month, with generous free trials. Other combinations available from the sales team.

RapidMiner

The tools from RapidMiner were always pitched first to the data scientists on the front lines. The core offering is a complete visual IDE for experimenting with various data flows to find the best insights. The product line now includes more automated solutions that can open up the process to more people in the enterprise through a simpler interface and a guided series of tools for cleaning the data and finding the best modeling solution. These can then be deployed to production lines. The company has also been expanding their cloud offerings with an AI Hub designed to simplify adoption.

Highlights:

• Great for data scientists who are working directly with the data and driving exploration
• Offers transparency for users who need to understand the reasoning behind predictions
• Collaboration between AI scientists and users is encouraged with Jupyter notebook driven AI Hub
• Strong support for Python-based open-source tooling
• Broad free tier provides RapidMiner Studio for early experimentation and educational programs
• Pricing for larger projects and production deployment available by request

SAP

Anyone who works in manufacturing knows SAP software. Its databases track our goods at all stages along the supply chain. So it should come as no surprise that they’ve invested heavily in developing a good tool for predictive analytics to enable enterprises to make smarter decisions about what may be coming next. The tool builds heavily on business intelligence and reporting by treating predictions as just another column in the analytics presentation. The information from the past informs the decisions about the future, mainly using a collection of machine learning routines that are highly automated. You don’t need to be an AI programmer to get them to run. Indeed, they’ve worked to create what they call “conversational analytics” that can provide useful insights to any manager who asks the question in a human language.

Highlights:

• Great for stacks that already rely on deep integration with SAP’s warehouse and supply chain management software
• Designed with low-code and no-code strategy to open analytics to all
• Part of a regular business intelligence process for consistency and simplicity
• Users can drill deeper by asking for context behind the predictions to understand how the AI made the decision.
• A free plan allows experimentation. Basic plans start at $36 per user, per month. More capable plans with more automation and integration available from the sales team.

SAS

One of the oldest statistics and business intelligence packages from SAS has grown stronger and more capable with age. Companies that need forecasting can produce forward-looking reports that depend on any mixture of statistics and machine learning algorithms, something SAS calls “composite AI.” The product line is broken into tools for basic exploration such as Visual Data Mining or Visual Forecasting. There are also some focused tools for specific industries such as the Anti-Money Laundering software designed to forecast potential compliance problems.

Highlights:

• A great collection of focused tools already optimized for specific industries such as banking
• Excellent merger of traditional statistics and modern machine learning
• Designed for both on-premises and cloud-based deployment
• Pricing depends heavily on the product choice and the usage

TIBCO

After data is gathered by various integration tools, TIBCO’s predictive analytics can start generating forecasts. The Data Science Studio is designed to enable teams to work together to create low-code and no-code analytics. More focused options are available for particular data sets. TIBCO Streaming, for instance, is optimized for creating real-time decisions from a time series of events. Spotfire creates dashboards by integrating location-based data with historical results. The tools work with the company’s larger product line designed to support data gathering, integration, and storage.

Highlights:

• Great for supporting a larger architecture for data management
• The predictive analytics integrates with several data movement and storage options
• Builds on a tradition of generating reports and business intelligence
• Machine learning and other AI options can improve accuracy
• Products are priced independently with a variety of different plans for cloud and on-premises options. Turn-key AWS instances begin at 99 cents per hour. Many options are priced by the sales team.