ChatterBox

Using Twitter Distribution Hubs

While Twitter provides a great platform for global information, how do you get to just the subset of information relevant to your local area. The answer is a system of Twitter distribution hubs (a.k.a. spoke-hub distribution paradigm). The idea is not a new one. Distribution hubs are used across the board for a multitude of purposes, including airline flight patterns, shipping, networking and many other reasons. According to Wikipedia, this distribution model is defined as:

“The hub-and-spoke distribution paradigm (or model or network) is a system of connections arranged like a chariot wheel, in which all traffic moves along spokes connected to the hub at the center. The model is commonly used in industry, in particular in transport, telecommunications and freight, as well as in distributed computing.”

Wikipedia uses this picture as the way to show how this works for airlines. So, in this example Denver serves as the hub for getting to and from place that do not have direct connections.

Twitter does not have a direct model for creating these types of relationships, but it does have a virtual way of setting up distribution points- search and hash tags. While Twitter’s search operators page at http://search.twitter.com/operators lists two operators that could apply (near and within), those work in the opposite direction. “Near” finds all Tweets sent from a specific area and “within” constrains the search to a specific distance away. This would be interesting for an aggregation approach, but since we are trying to distribute messages to the spokes and not consolidate messages from the spokes, we need a different approach.

Fortunately, a standard has emerged around using hash tags (#some_topic) which allow people to find things around a common topic without having to follow a specific handle to be notified of the information. You can then use Twitter search to find all messages related to that topic. While not as convenient as near and within, a hash tag model could be applied to describe messages destined for specific areas. You could also tack on other tags to further refine the distribution path. Since these tags are currently freeform, a standard would be needed to ensure that everyone leveraged common formats for the tags. The critical piece here is that Twitter searches can be made into RSS feeds. So, I can easily set up a search for a specific hash tag and am notified when a new item matches that search.

So, how would this work. The first step would be to establish a set of hub information providers. Next, you would establish a set of Twitter searches based upon a hash tag specific to each spoke. Using a tool like RSS to Twitter (no affiliation) you could feed the results of those search to Twitter handles specifically set up for the spoke location. Users could follow the spoke to then be notified of the specific items of interest.

There are a multitude of sample uses for this type of a model (e.g. local advertising), but a more interesting use could be to create a localized emergency broadcast system for things like Amber alerts, weather emergencies, or other types of local notifications. The system would look something like this:

An organization like the Homeland Security Department would operate as the hub sending a stream of messages tagged appropriately for distribution. The messages could embed links and pictures for longer text or photos of missing children, persons of interest, etc. Local agencies would then follow the spoke accounts to get the subset of information relevant for that specific area.

The benefits of this approach include the following:

• Notification logic can be centralized to a single point of access which simplifies distribution
• Instant global reach
• Little development is required to integrate with Twitter
• Twitter is easily accessible by mobile devices
• Content can easily by filtered down to a specific interest areas

There are a few drawbacks to this approach. The main drawback is the stability of Twitter itself. Without a reliable mechanism for distributing these messages the usefulness of leveraging Twitter in this fashion is extremely limited. The second drawback is that spoke-to-spoke (or point-to-point) messages can not easily be sent directly since all communication flows through the hub. Fortunately, Twitter provides an easy mechanism to forward messages. A spoke could send a message back to the hub with a hash tag for another location. The hub would then re-tweet the message and it would be picked up by the other spoke. Another thing to keep in mind is that traffic would be bottlenecked by the hub account. So, if things like API limits were not lifted, message distribution could be delayed. Additionally, without a standard for hash tags, coordination across the network would incur a high level of overhead.

Overall, this is another idea to keep in mind when looking at notification systems. Email and automated phone calling have already greatly improved response times and awareness of incidents. Leveraging Twitter in this model is yet another improvement.

Todd

Posted by: Todd Clayton  |  April 21, 2009  |  categories: Social Media  |  tags: Distribution, Social Media, Twitter