Datavisualization with Stata: How to adapt APIs and take benefit of HTML5/canvas to create dynamic charts and maps with Stata

Dr. Thomas Roca, Agence Française de Développement - ERS/ RED - Research Department

Datavisualization is a burgeoning field at the crossroads of design, computer science and statistics. Using
HTML and applying datavisualization technique allows producing elegant and insightful representations with
Stata. Nevertheless, creating original dataviz from an empty page is rough, it requires specific programming
Knowledge. Thankfully, many pieces of code have been developed under GNU/GPL or Apache license, accessible
for free (Google APIs are the most famous ones).
However, these pieces of code need to be adapted; many options are available to modify the display (texts,
legends, labels, colours, size, type of representation, etc.) Besides, HTML5 now embeds a powerful graphic and
drawing engine: canvas, which can be mobilized to represent any data, starting from scratch. As underlying
datasets need to be reshaped and organized, a fair amount of programming becomes necessary; thus, Stata users
possess a great asset to fill in and format HTML and Java content.
Specific programmes will be discussed:

gmap package creates heat maps using Googles API.

himap package builds two layers pie charts supported by Highcharts API.

scatter3D package builds a 3D scatter plot using HTML5's canvas Xpress API.

Dynamic Scatter Plot, fuelled with HTML5/Canvas, can represent 3 variables, and displays additional
information, scrolling over with the mouse. Furthermore, two Dynamic Scatter Plots can be superimposed to
facilitate comparisons.

To help grasp the possibility offered by Stata in web programming, the Country Dashboards project will be
presented. It shows how to create a comprehensive web portal, embedding hundreds of dataviz and web pages
built with Stata.

About the speaker
Thomas Roca is a researcher and statistician at the French Development Agency (AFD), Research department. Thomas is developing AFD’s research program covering well-being, Human development & alternative welfare indicators, including Big Data for Development. Thomas’s filed of work covers also datavisualization and programming. Within this programme, partnerships have been built with UNDP, UNU-merit, OECD’s , ODI-MIT-mediaLab and Université Paris I – ESSEC. As statistician, Thomas was in charge of the Institutional Profiles Database 2012 (IPD), a global perception survey, probing institutions and governance quality. Before joining AFD in 2011, Thomas was a cooperation officer at the French ministry of foreign affairs -French Embassy in Hungary. He taught Economics in Hungary, Vietnam, and Kurdistan as a lecturer for Université de Picardie. During his PhD, probing governance measurement, Thomas interned twice UNDP: in Morocco within the governance program (2006) and in New York, Human Development Report Office (2010).

Venue: Conference Rooom

Date: 30 October 2014

Time: 12:30 - 13:30


UNU-MERIT