BBVA API Market
Predictive analytics is “a business game changer”. That would be the perfect summary of the latest Forrester Research report with its annual classification of the best predictive analytics solutions for companies. There are no more excuses. All the companies in the world have such advanced and easy to use data analytics solutions available that not knowing what will happen within six months is a mortal sin.
The major development of advanced analytics is the ease with which high-ranking corporate executives can make predictions and understand the future of their teams, independently from the work already being carried out by other profiles such as data scientists and data analysts.
This explosion of accessibility involves two key points:
1. The use of APIs: companies in sectors such as finance, retail, or energy use application programming interfaces to construct predictive models and extract value from the data in order to:
– Draw conclusions from the data before making decisions.
– Predict customer behavior in order to adjust the offering and prices; and discover their opinion on products or services.
– Know how to increase productivity and performance.
– Prevent or detect fraud.
2. The emergence of the Predictive Model Markup Language (PMML): this is an XML text markup language developed by the Data Mining Group (DMG). It is a standard language used to represent predictive models. It allows the same solution to be shared by different PMML-compatible applications.
The solutions of the best-in-market providers
1. IBM: this is one of the key players in the predictive analytics market. The company has a number of solutions:
– Customer analytics: IBM has a tool to predict levels of customer satisfaction or dissatisfaction, retain them and increase revenue. The tool allows companies to create personalized offerings in different channels, predict when customers are about to change supplier, detect market trends in social networks, use cross-selling techniques…
– Operational analytics: the idea is to provide companies with the tools to evaluate operating costs, speed, flexibility and quality through data collection, storage and analysis; and to find value in them in real time.
– Predictive analytics for big data: most companies have a large volume of data. The problem is that these data are unstructured or semi-structured and decision-making in this scenario is impossible. The key is to have tools that organize the data, extract relations and make projections without the need for high levels of technical knowledge.
This is achieved with tools that combine unstructured data simply, prepare visual and plain language summaries so that the business analyst (not data analyst) understands the value of the information, and establish demand projections and personal customer profiles. The IBM predictive solutions are open and support Hadoop technology.
– Threat and fraud analytics: these predictive models help anticipate any threat of fraud in companies. The tools detect unusual patterns in information using data mining and analysis techniques.
2. SAS: SAS Visual Statistics is a platform for analyzing a large number of data stored in Hadoop simply. The tool’s user interface allows any executive to drag and drop variables in order to create dashboards.
The SAS tool has some interesting analytics features:
– SAS Analytics Visual Explorer: creates interactive models based on multiple variables. All you have to do to generate visualizations with these variables is drag and drop: bar charts, histograms, box plots, heat maps, geographic maps, bubbles, dispersion diagrams… It’s all very visual.
– Interactive descriptive modeling techniques.
– Construction of predictive models using techniques such as linear regression, generalized linear models, logistic regression and classification trees.
– Comparison of models with the creation of summaries: lift charts, ROC curves, concordance statistics and misclassification tables.
3. SAP: this company has one of the best-known solutions on the market, SAP HANA, which combines database and application platform capabilities in a single in-memory solution. In the case of the creation of predictive models, the platform provides libraries for text processing, spatial processing and business analytics in a single architecture. What are its advantages? It is capable of processing a large number of data in real time with no latency. SAP places a great deal of emphasis on the speed and scalability of HANA
– Up to 3.19 billion symbol scans per second per core.
– SAP HANA set a new record for the world’s largest data warehouse: 12.1 petabytes of data, four times larger than the prior record.
4. Oracle: Oracle Advanced Analytics is a tool that extends the Oracle Database into a single platform with the addition of the advanced data analysis of Oracle R Enterprise and Oracle Data Mining. The company’s solution provides data analysis in real time in key subjects such as customer churn prediction, product recommendations and fraud alerting.
– Oracle R Enterprise provides libraries in R, the programming language used in statistics, that can analyze data contained in any database.
– Oracle Data Mining provides powerful data mining algorithms to build, evaluate, share and deploy predictive analytics models.
5. RapidMiner: RapidMiner Studio is a platform for predictive analytics, machine learning, data mining and business analytics. It allows the data loading, transformation and modeling of large volumes of data from sources such as Excel, Access, Oracle, IBM DB2, Microsoft SQL, SAP Sybase, Ingres, MySQL, Postgres, SPSS, dBase, and any plain text source. In addition, companies can integrate their own data algorithms into RapidMiner through its open extension APIs.
6. Alteryx: Alteryx 7.1 is a platform that provides data access, management and predictive analytics in the same tool. It is a solution that integrates all the functionalities of the R programming language for statistical study, but without the need for the user to have advanced statistical knowledge.
Thanks to Alteryx, a company can handle large volumes of data and interpret them using spatial analysis. In addition, as with other predictive analytics tools, users can drag and drop variables to create their own models.
7. Microsoft: the solutions provided by SQL Server, the database platform of the Redmond-based company, allow clients to obtain valuable information from predictive data analytics through the integration of in-memory and high-performance technology with Online Transaction Processing.
SQL Server uses a set of tools to implement and administer databases in the cloud and in local environments, allowing its clients to design predictive analytics solutions in two ways:
– The business analyst can use data mining techniques with tools such as Excel with Data Mining Add-ins to derive patterns, create visuals and graphics with the data and generate visual summaries.
– Developers can use SQL Server Management Studio to create ad-hoc data mining and predictive analytics solutions.
8. KNIME: KNIME Analytics Platform is a data mining platform for the creation of visual predictive models. It is a solution based on Eclipse and is written in the Java programming language. With this tool, you can for example:
– Visualize data in histograms, maps…
– Create statistical models: decision-making trees, regressions…
– Draw up personalized reports.
– Incorporate written functionalities in Python or R.
It is an open code solution, under a GPL (General Public License). The tool can also be combined with commercial solutions. In any event, any user can download the free platform and begin to use it without any problems.
Follow us on @BBVAAPIMarket
Real-time payments have become one of the most noteworthy innovations in the financial industry. Their growth in recent years has been significant thanks to the possibilities they offer companies, especially in customer relations.
Open finance has become one of the main drivers of digital financial change worldwide. Its implications go far beyond those of open banking and may serve to change the current financial paradigm in some regions such as Latin America or Africa.
Lack of confidence in the payment platform and slow processing are the two factors that most influence the abandonment of purchases from ecommerce. Two problems that can be solved thanks to APIs.