Monday 5 October 2009

"How vague is it?", a challenge for neogeographers


People usually don't like caricatures (especially those who are caricatured!), but sometimes they are useful to sketch quickly and efficiently a situation and highlight a problem to solve. So let’s try to caricaturize neogeography, one hot topic of the moment, in order to uncover interesting challenges … (and of course no offense to anybody, neo- or paleo :-) ).

Let’s assume Mr.X. He’s neogeographer: he’s young, freshly shaved, technologically ultra-literate and he’s convinced that the Internet is nothing less than the nervous system of the planet. He likes mashups, he likes mapping APIs, and he considers that the wisdom of the crowds provides him with any information he would really needs (‘if it’s not on Twitter, it doesn’t exist’, he uses to say). He makes pretty beautiful maps; usually, they are not very useful, but so appealing!

And then comes Mr.Y, the paleogeographer, with a grey beard and an outdated desktop GIS package. He trusts only data he has paid for (so that there is a written trace of the provider’s liability), and carefully references in a proper metadata record the information he derives from his sophisticated geoprocessing scripts(he has a thousands of scripts written in an obscure language, that can derive everything he needs to know about the planet earth). Using his trusted data and models, he provides scientific evidence that nobody would dare to discuss on. The paper maps he produces are often ugly and outdated, ‘but at least they’re 100% right’, he uses to say.

The thing is that Mr.X will most likely never extract some real meaning of the extraordinary mass of information he’s playing with. And Mr.Y will never make he’s great models benefit from this abundant mass of geo-information Internet can provide. So now, OK, I told you it was a caricature and this statement can appear exaggerated, but here’s my point: I would like to see more added value to geodata out there (who rememers useful concepts like ‘feature overlay analysis’, ‘density’, ‘proximity’, ‘map algebra’, etc.?) and I would like to see geodata provided by non-specialists (i.e. Volunteered Geographic Information – or VGI ) more often considered as a novel source of information for our earth monitoring processes.

Building bridges between casual- and expert geography, this was the (wide, wide picture of) the motivation behind the paper I presented last June in GSDI11 conference. The idea was: if we can assess ‘how vague’ is the information provided by citizen when acting as geo-sensors, it’s a good start to use this data ‘for what it is’, recognising its value but also its limits and build some robust quality control procedures on top of it. The underlying topic of fuzzy geography is not new at all; some eminent scientists worked on it since decades. But maybe it worth it to re-examine it with a fresh perspective, like a neographer would hopefully do …

Thursday 19 February 2009

User-Created Geographic Information Roundup

These days, we see flourishing on the web numerous sytems that allows people to create contents with a geographic component. We can call their output 'Volunteered Geographic Information' (scientific trend) or 'crowdsourced geotdata' (buisness trend).

Here follows some examples of such systems:

  •  OpenStreetMap (http://www.openstreetmap.org/) is one of the most known one. OpenStreetMap is a free editable general purpose street map created by collaborative methods. It aims to cover the whole world. At time of writing, OpenSteetMap contained more than 22 millions of kilometres of roads, covenering (heterogeneously) 114 countries on the 5 continents.
  • Wikimapia (http://wikimapia.org/) aims to describe places all over the world. Inspired by Wikipedia, this project focuses only on articles that can have a physical location (i.e. about towns, cities, lakes, regions, etc. and not about persons, technologies, etc.), and offers a map interface to browse them. Notice that in parallel, several projects inside Wikipediapromotes the addition of geographical coordinates to articles, when relevant.
  • Many platforms specialised in user-generated media are offering Geo-tagging functionalities. In consequence, a huge amount of pictures and videos can be browsed in their spatial context. For example, flickr contains an archive of more than 90 000 000 picturesuploaded by a huge community of photographer , and Panoramio a similar platform offers16 000 000 geotagged pictures. Youtube, the famous video-sharing platform also supports interesting geographic functionalities: every uploaded video can be geotagged using a simple map widget, its API allows to easily search geo-tagged videos by location and to display it as a KML layer in any web map client.
  • Some sites offer possibility to exanche real-life points of interests (hotels, restaurants, etc.) like http://platial.com/ or http://poifriend.com/ the latter offers a kind of 'community' features (i.e. share your points with your freinds), but when will we see a full integration of such features with our prefered Social network? (e.g. Facebook, MySpace, etc.). I think it is vital for such infomation sharing (otherwise there is no added value, just search any phone directory). 
  •  MapMyFitness (http://www.mapmyfitness.com) targets a specific community that deals a lot with geographic information for their leisure. This  platform allows outdoor sports practitioners (cyclists, runners, hikers, tri-athletes, etc.) to map their favourite routes. In addition, this platform offers community features like groups creation, sharing of routes, rating and comment other’s routes).
  •  On a very different register, Ushahidi aims at crowdsourcing crisis information http://www.ushahidi.com/. The system allows victims to testimony (ushahidi means ‘testimony’ in Swahili) in real time about the evolution of the crisis on the field around them, by providing geo-referenced messages by e-mail, SMS or via the web. Ushahidi has been deployed during the recent troubles in Kenya, in Democratic Republic of Congo, and in the Gaza stripe.

These examples illustrate how User-Created Geographic Information can concern vast amounts of data, and find applications in various domains. In addition, it becomes always more ubiquitous, with, the possibility to geotag Blog posts, or to generate geotagged messages, photos or videos directly from GPS-enabled smart phone.

I see a lot of potential buisness around User-Created Geographic Information and related technologies in the near future, as it fits very well to the emerging characteristics of the web platform: social dimension and mobile applications. The omnipresent (and omnipotent?) Google has understood it very well with the launch of Google Latitude. But for once (I'm a fond of Google analytics, Google Calendar, Google Docs, Google Spreadsheets, Google Maps, ...) I don't like their appoach as I find it too tech-dependant and too intrusive. Maybe a topic for a future post?

Wednesday 4 February 2009

Web Map Clients: Flex vs AJAX

Without being aware of it, ESRI recently gave us an interesting point of comparison between to interesting technologies for web map clients. Indeed, on its ‘community’ website (http://resources.esri.com/) the Redlands company released  nearly at the same time two web map client code samples that everybody can upload and test

Flex client: http://resources.esri.com/arcgisserver/apis/flex/index.cfm?fa=codeGalleryDetails&scriptID=15905

Javascript (AJAX): http://resources.esri.com/arcgisserver/apis/javascript/arcgis/index.cfm?fa=codeGalleryDetails&scriptID=15987

Unfortunately, I didn't manage to find online working demos, but I tested in on my 'localhost', and I can tell you that they are very similar in terms of user experience!

They look exactly the same, and behave also in a very similar way (try for example to add the earthquake GeoRSS,  or to draw your own vectors) … but behind the scenes, technology is drastically different!

So what to choose if you need a Web Map client technology? I would clearly advice AJAX because:

  • Flex is a proprietary technology (it is a kind of ‘new branding’ for the great old Flash, owned by Adobe), and runs in compiled mode (.swf files are binary) so it is very bad for Open Source projects;
  • AJAX is a technology that has been built to provide interaction with underlying web services (or any form of XML or JSON based communication), while Flex is initially a vector drawing technology;
  • You can develop in AJAX using any Text editor. If you need more elaborated IDE, several open source software are available (like Eclipse with the Aptana plugin). In contrary, advanced Flex development requires Flex Builder (215 euro on Adobe online shop).
  • As a consequence of the previous point (and a lot more important than a simple license pricing issue), AJAX is supported by a huge community of developers while Flex remains more confidential. Many AJAX open source projects like OpenLayers, MapFish or ka-map are available, with a lot of code samples, forums, etc. that really improves productivity of developers using them.

Among those Open Source web map client technologies, MapFish (http://www.mapfish.org/) is one of the most promising projects, I think. It puts together a Javascipt library dedicated to web mapping (OpenLayers) and a library efficient GUI design, and powerful AJAX request management (ExtJS). As a consequence, it is very powerful in terms of built-in OGC support (WMS, WMC WFS-T, KML, GeoRSS, etc.) and mapping features (deal with projections, map caches, Google maps layers, zoom, pan, vector drawings, etc.), as well as in terms of client-server communication and capacity of building without effort nice looking web user interfaces.

Also of interest, the beta client developed by ESRI (them again!) with the Geoportal Toolkit 9.3: http://keel.esri.com/Portal (the ‘LAUNCH MAP VIEWER’ link is on the upper right). But even if based on AJAX, this is a proprietary technology, so definitely not a solution that cannot fit to any project.