Titulo Estágio
Visualizing transactional patterns to support fraud detection
Áreas de especialidade
Sistemas Inteligentes
Local do Estágio
Portugal
Enquadramento
Feedzai’s risk analysts face the challenge of identifying data patterns and comparing real-time events with historical client information. This non-linear, iterative data exploration of tabular data makes it diffi cult and time-consuming to analyze information. When reviewing fraud alerts, analysts must rapidly explore sometimes hundreds or thousands of data points within a few minutes. To add to this already complex task, they need to shift between aggregated data views and granular explorations of concepts like geographical information, device data, triggered rules, transaction amounts, and behavioral patterns. Combining these diverse sources is crucial for analysts to make accurate and timely decisions, highlighting the importance of data visualization in their workfl ow.
Objetivo
The goal of this project is to explore ways of assisting fraud analysts in their review by exploring, validating, and prototyping data visualization techniques specifi cally tailored for the detection of fraudulent activity. You’ll be faced with the challenge of understanding and visualizing complex data with multiple dimensions (for example transaction amount, location, time, network connections) to provide these types of view to analysts.
In this project you’ll have the opportunity to engage on the different tasks required to produce data visualizations that make a difference for our users. By applying user research techniques to understand user needs, creating mockups of the visualizations/interfaces that may respond to those needs, prototyping the solutions proposed, and validating them with users. Throughout this process, you’ll be supported by senior Data Visualization researchers that will help you in each step of the way.
You’ll also have the opportunity to engage in the production of scientifi c papers based on your work.
Plano de Trabalhos - Semestre 1
Literature Review: Explore academic literature related to data visualization, fraud detection techniques, and user research methodologies. ● Technology Planning: Determine the technologies and tools to be used for developing and prototyping data visualization solutions.The team uses React, Typescript, visx, d3, Three.js , among others. ● Understand Feedzai: Learn how data visualization is currently used within Feedzai. ● User Journey Understanding: Familiarize yourself with the "user journey" of risk analysts at Feedzai, focusing on their data exploration and analysis processes. ● User Research: Conduct interviews and/or other user research activities with Feedzai users to understand their specifi c needs and challenges related to data visualization for fraud detection.
Plano de Trabalhos - Semestre 2
Visualization Design: Create mockups and designs for potential data visualization solutions based on the research and requirements gathered. ● Prototyping: Develop interactive prototypes of the proposed data visualization tools and interfaces. ● User Testing and Validation: Conduct user testing sessions with Feedzai users, to gather feedback and validate the effectiveness of the prototypes. ● Thesis Writing: Write the dissertation/thesis document, documenting the research process, fi ndings, and developed solutions.
Condições
● This is a remote position within Portugal for a paid internship (remote, hybrid or from one of our offi ces in Portugal). The team welcomes applications from students excited about improving user experiences with data visualization and keen on contributing to the advancement of the fi eld by producing academic papers and eventually pursuing a PhD in the fi eld of data visualization. ● You’ll have the opportunity to engage with real users of our product and validate your proposals with them.
Observações
Past work from the Data Visualization team at Feedzai:
“Show Me What’s Wrong!”: Combining Charts and Text to Guide Data Analysis
AutoVizuA11y: A Tool to Automate Screen Reader Accessibility in Charts
A case study on implementing screen reader accessibility in dynamic visualizations
Supporting visual investigation of data distribution shifts by data scientists
Orientador
Rita Marques Costa
rita.costa@feedzai.com 📩