Xenon APAC is a supporting application for algorithmic trading platform. It helps maintain client reference data for Asia Pacific region. The application is being used by trading desk while processing trading signals.

The user interface has been developed using Python over Django 2.0. Jinja2 template engine has been used under MTV architecture. Automated unit tests have been written in Selenium Robot framework.

Apache Hadoop is being used to process data along with Hive which is being used to support HSQL for mining purpose. Sqoop is in place to transfer structured data (from Oracle) to Hive/HDFS.

The application’s analytical features have been developed using opensource QuantLib and Python libraries. QuantLib is being used for developing modelling features like PDEs, yield curve models, Monte Carlo, VAR etc. Python libraries – NumPy, SciPy, Pandas and Scikit-Learn have also been used.

The project is evaluation of various data science tools and practices. The objective of the project is to develop standard data science process within organization.

The project consists of various case studies and POCs. The machine learning model, statistical modelling and technical evaluation has been done using Fannie Mae Loan Performance, Freddie Mac loan Performance, Primary Mortgage Market Survey dataset. The POC has

been developed to showcase a reliable Loan Predictive model. It involves further development of pre-payment model.

Abnormal Trading Pattern (ATP) is a hybrid implementation of Algorithmic Trading and Data Analytics/Data Science. Each Trading Activity generates alert(s). The generated alerts are analyzed statistically and processed further. Underlaying algorithm ensures the accuracy of generated reports. User can also export the analyzed data in spreadsheet.

Predictive models have been developed using various statistical techniques like Linear Regression, Multiple Regression, Logistic Regression methods and Cluster analysis etc. Machine Learning algorithms have been developed using concepts of k-NN and Naive Bayes functions.

Web has been developed using AngularJs, JavaScript and Angular Material. D3.js has been used for data visualization and presentation. Various statistical charts were used to visualize and present data like Bar Charts, Line Charts, Scatter Plot and Donut Chart. Libraries (D3.js, ggplot and matplotlib) for Angular, R and Python have been used. Oracle 12.0 is being used as DB which is being accessed using Oracle adapter.

DevOps is a set of best practices that targets Continuous Delivery through a set of advanced tools during ALM (Application Lifecycle Management) by using techniques such as continuous integration, automated testing and continuous deployment.

The purpose of DevOps is to set up a process for Continuous integration of projects using TeamCity along with suitable tools for Code review, calculation of Code metrics, deployment and security review that are relevant to the technologies used. Visual feedback of the results at different stages of the process is available on a Dashboard. The metrics can be used as parameters to judge the quality of the code developed and suitable remedial action can be taken when required.

The application also used Tableau and NodeXL as Data Analytics tools and to analyze data and user pattern. It further generates report presentable for leadership.