writing code to reveal data
overview
We write software to visually represent data,
trying to reveal the information that is hidden in it.
A visual representation of data lets you discover more about data than just reading the numbers. A visual representation of data gives you the big picture in big data.
case studies
A partial and heterogeneous list of projects:
Public mobility through WIFI traffic data
A collaboration with the University of Bergamo (Università degli Studi di Bergamo)
Data: All logs of a public network of WIFI Access Points
Goal: Infer informations about mobility in that area
Output: Videos for broadcasting, rendered through a dedicated rendering engine created for the project
3D WebGL navigation engine SVAR
A collaboration with the University of Bergamo (Università degli Studi di Bergamo)
Data: Indoor Positioning System realtime coordinates and architectural DXF files for all of the buildings
Goal: Create a navigation system for indoor public spaces
Output: A PWA WebGL web application for PCs and smartphones, with Indoor Positioning System and 3D interactive navigation capabilities
Simulating evolution of a community of individuals
An interactive visual representation of the evolution of a community of individuals.
Data: Some food sources, and a population of hundreds or thousands of individuals, each one with a different and unique color (DNA), fighting for food
Goal: Simulate how the community evolves in time, and highlight what's the color (DNA) with highest survival chance
Output: A graphic simulator, all written in pure javascript.
our process
Depending on data size and exitamed cpu usage, we must provision hardware requirements for the project, and we do it before starting to write code.
Data could be numbers, words, relations, pictures, anything... we dedicate appropriate time analyzing data sources, to understand what's the significant information that must be highlighted in the visual representation.
Usually a lot of data is considered noise and must be skipped, and in case of big data some sort of data filtering/reduction is mandatory to keep cpu time under control. To address these tasks we write custom cleaning and normalization algorithms that start from original data and generate cleaned/normalized/reduced data ready for the rendering process.
We support output on any media. From videos for broadcasting to PDF for newspapers and magazines, to interactive for web. Based on media type and on data insights, we write a custom rendering engine to create the graphic visualization that better represents data.
For customers that need to repeat the rendering of the graphics on a scheduled basis, for example because the data is updated periodically, we can automate the process, and deliver updated graphics via FTP to the customer in a totally automated way.
Graphics engines we create are easily integrable with customer infrastructure, thanks to customizable REST APIs.
information