Chapter Synopses
The following sections briefly outline each chapter of the project report. You can download the full report here.
Chapter 1 - Introduction
This project investigates the process of building an information browser for rich, multimedia data that has a geographical component. This includes building a database for retrieving coordinates given location-names. We also cover the problem of extracting and tidying existing data which contains some sort of geographical reference. Sometimes this is explicitly defined, and sometimes we have to apply some informed guesswork to establish where the information is referring to. Finally, we investigate various visualisation tools that allow us to develop powerful applications which can display and explore the geographical information that's available.
Chapter 2 - Geopgraphic Name Searches
In analysing data sources with a geographical component, we are often not provided with
exact coordinates coupled with every place name. For example, one may be able to obtain a data
feed containing a list of job-vacancies in the UK. The provider of this feed is highly unlikely
to tag each location with a latitude and longitude value. With this in mind, we constructed and
optimised a database of place names and coordinate values which was sourced from various US
Government agencies. Algorithms were also built to improve the ranking of places in search results.
This was particularly important when more than one location had the same name. You can experiment with
the geocoding tools created in this chapter at our tools page.
Chapter 3 - Collecting & Extracting Data
In this chapter, we investigates methods of collecting or extracting data that is of geographical interest. The BBC make a lot of interesting information available to us through their Backstage website in well-structured XML or RSS feeds. Other information on the web is often not so well formalised - we were interested in creating maps from plain text documents and HTML files where the information is surrounded by images and styles which are irrelevant to the map. Within the text, we built methods of detecting which words are place-names (in order to position the map point correctly). We also created methods for accessing the collected data within any application. Here we introduce the Map Builder Markup Language, or MBML, which was later used in Chapter 4 by the main visualisation tools created.
Chapter 4 - Visualisation Tools
In this chapter, we explore the various ways of creating a data browser specifically for data with a geographical component. We mostly investigate Google Maps - a public service offering a web-based interface and an API for developers. This API is used to build the Map Builder which, when coupled with the geocoding service outlined in Chapter 2 and the MBML format introduced in Chapter 3, is capable of producing maps from any data set. Finally we look at the opportunities within the mathematics package Matlab which allow us to analyse the geographical data and display it in a simpler way. We produce maps which allow the display of numeric data in interesting ways, and also generate animated maps - allowing us to visualise geographic data that exists in a time-series. You can experiment with the many of the visualisation tools at our tools page, and see some of the animated time-series maps at the videos page.
