Programmers amongst you might want to look at the code, but here's a quick overview.

LifeMapper takes a list of items, which you rate for passion and experience. How much you love something, and how much experience you have with it.

Lifemapper uses Wikipedia to find how related two concepts are. Say you choose "cheese" and "maths", LifeMapper will start at cheese, and finds the smallest number of clicks needed to gets to maths. It then counts from maths to cheese, and uses the average value to show how related two concepts are.

This alone is quite a crude method of finding how similar things are, but as you'll see later, as more connections are found, the accuracy improves.

As you add items, the system builds a network of how all your items link to each other, finding how related each item is to every other item. This can be imagined as a network, such as the one above, where very related items are closer. Ignore the color for now.

It might be the case (in fact, rather likely) that you can't make a 2D network where all the lines are the right length, so if you wanted to make a real model, you might have to use 3D, or higher dimentions. But luckily we don't need to draw it. :)

Ok, so now you've got a network of your items. Rating items (the nodes of the network) you add a value to that node. (Shown in red above). For nodes you haven't rated, the system will estimate a value based on the values around it weighted by the distance to each.

For estimates "not from your lifepoints" the system builds a network of all inputted user points (a few hundred now), and uses that to estimate. In a perfect system it would use the whole of wikipedia... but that's millions of links... and a very slow system.

LifeMapper takes a list of items, which you rate for passion and experience. How much you love something, and how much experience you have with it.

Lifemapper uses Wikipedia to find how related two concepts are. Say you choose "cheese" and "maths", LifeMapper will start at cheese, and finds the smallest number of clicks needed to gets to maths. It then counts from maths to cheese, and uses the average value to show how related two concepts are.

This alone is quite a crude method of finding how similar things are, but as you'll see later, as more connections are found, the accuracy improves.

As you add items, the system builds a network of how all your items link to each other, finding how related each item is to every other item. This can be imagined as a network, such as the one above, where very related items are closer. Ignore the color for now.

It might be the case (in fact, rather likely) that you can't make a 2D network where all the lines are the right length, so if you wanted to make a real model, you might have to use 3D, or higher dimentions. But luckily we don't need to draw it. :)

Ok, so now you've got a network of your items. Rating items (the nodes of the network) you add a value to that node. (Shown in red above). For nodes you haven't rated, the system will estimate a value based on the values around it weighted by the distance to each.

For estimates "not from your lifepoints" the system builds a network of all inputted user points (a few hundred now), and uses that to estimate. In a perfect system it would use the whole of wikipedia... but that's millions of links... and a very slow system.