This is a first try of a diagram for the social database that is going to store all the location data for the Talking-Points project. The database has two main goals:
1) make it as easy, accessible and usable as possible to contribute to the database
2) provide relevant and reliable information for each location
The format of the diagram is similar to an ER diagram but at this point only intended to show the concept of the database and not yet the actual tables that it will consist of.
The basic concept of the database is to differentiate between two kinds of location information. On the one hand (marked with the red ellipse) the core location information and on the other hand additional information.
The core information is supposed to include the most relevant, straightforward and reliable information about the location. This includes a meaningful title, a short description and location metadata such as coordinates, address, official URL and so on. This information is created when a location gets tagged with an RFID tag. Important about this information is that it needs to be objective and reliable. We are not 100% sure how we will ensure this, but there will be some form of checking and moderation. You can think of it a little bit as stable versions for Wikipedia.
The additional information is where the social part comes into play. Once the user has some basic information about the location he might want to know more. In this case we would like to offer different directions/types of additional information:
- One direction is going to be opinions. Here the users can rate the location on a 5-star scale and state their opinion.
- A second direction would be current information about the location. This could be opening times, news stories related to the location or also current special offers in case the location is a business.
- The last direction would additional general information about the location. This information would be stored in maintained in a wiki page so that every user could add what they think is relevant concerning the location. One topic that would probably be interesting for many locations, would be something about its history.
In addition to that you will also be able to add links and tags to the locations. The tags could be used to only get notified about certain locations that are tagged with for example “blind” or “tourist”.
To see which information will actually be presented to the user of the mobile device I have marked those with the green ring. Anything else will be used only indirectly for this part.
One really major thing that we still need to address are incentives for putting information into the database. The design should already be fairly easy and usable, so that barriers to use it are low. But lack of barriers is not enough, there also need to be positive incentives. There is the incentive that you will help blind people by maintaining location information in this database, but how could user make personal use of that?
As with anything on this blog, any feedback that you have concerning our database design is very welcome and will be highly appreciated.
Great post Jakob! Very informative and easy to understand how you thought about the structure of this db.
Have you thought about repurposing some of Arborwiki for this project?
http://arborwiki.org
It has the right level of detail about the city (1000+ pages), and many of them are geocoded with locations already; it has reviews and information about businesses, and some level of street and pedestrian level navigation organized.
That could collapse the design down quite a bit:
* where am i?
* locate pages in the wiki that are near that spot
* list them by name to the user
* select one by name, and then read it to the user
As to the “where am I?” question, an additional data source you might use beyond RFID is to triangulate location based on the SSID of available wifi access points. In downtown Ann Arbor you can use the Wireless Washtenaw access points as where-am-i beacons, and then use cafes as wayfinding tools to go from place to place. No need really then to deploy special RFIC chips when the wifi infrastructure already gives you transmissions you can tune into.
Hey Ed,
thank you for your comment. An important point on our plan for the database was to incorporate as much external location information as possible. Source that we considered were for example Wikimapia and also the Arborwiki (although so far I only knew it existed and not many details). So we would interested in talking about how we could make the two systems work together.
In terms of the “where am I?” question we actually abandoned RFID as the technical focus of our design. Our prototype will now probably be based on Bluetooth. But we will keep on looking at other possible technologies (GPS, 802.11, Wifi triangulation, ….) and try to keep our design as open as possible so that the “where am I?” question could be answered by multiple methods. And actually the question in our design is more like “what am I passing right now?” or “what am i standing in front of?”. Primarily, we don’t want to give to much information to the user about what is near, but only the one location that he is close to at this point.
In any case we are very much looking forward to cooperate with your project.
Thanks! Here’s some more information …
There’s an existing “where am I” application database available throgh a couple of existing tools, some of which are mostly closed systems, and some that have APIs.
Plazes does the access point to location mapping, and there’s pretty good existing coverage of Ann Arbor.
Google Maps for Blackberry does cell tower location, but it’s super iprecise, so you know where you are within a few blocks and not much better.
The Getty Museum did a geolocation system using wifi and was really unhappy with the results – here’s a survey article with the state of the art 2005 results in museum applications:
http://www.archimuse.com/publishing/ichim05/Proctor.pdf
You could look at Yellow Arrow which uses stickers to be the access point IDs:
http://yellowarrow.net/index2.php
Happy to think about this more, there’s quite a bit of prior art to be inspired by and to improve on.