Travel planning is something that we always begin a few days before embarking on a trip, and while I have found it interesting, it is also very time consuming with loads of information available on the Internet. One cannot really decide if a certain place or activity is what one would like to do, and finding this is probably an interesting problem to solve using technology.
I have been thinking about it for a while and thought that with some user input, and all the information freely available about different cities, the places to visit, and types of activities they are most popular for, one should be able to automatically generate itineraries by mining this large information base. I came across a very cool website called Plnnr, which does exactly this. They have a really nice interface, with a few cities to choose from, and the itineraries that are generated are based on user preferences such as family trip, how intensive it should be, dates etc. Integrate this with hotel information and you have a pre-planned holiday in seconds. Going forward, one could probably add flight recommendations, weather related information to change plans on the fly, estimate of how much the entire trip would cost based on the recommendations, notifications about deals in the neighborhood, feeds from friends who have been to these places, etc but its very useful as of today in its current form.