So true. Google Calendar should rule this space, and it drives me nuts that deciding to attend an event usually involves multiple rounds of copying/pasting and manual editing when the amount of data involved is so trivial and well-specified.
On a broader level, there's a huge disruptive opportunity for a time-based search engine. When things happen is often very, very important, but gathering, collating, and connecting this information is painfully difficult. Take a simple example: Federal Reserve Board meetings. There are of tremendous importance to economic analysts, and worth knowing about for any politically aware person. Here is the calendar page: http://www.federalreserve.gov/monetarypolicy/fomccalendars.h... There isn't even an RSS feed. On a more general level, there isn't any obvious tool for browsing recurring events to look at their periodicity. To stick with the economic context (because it's one that matters rather a lot), consider the quarterly earnings and filings of public companies. As someone who likes DSP, it's perplexing to me that I can't add and subtract periodic data about companies' year-on-year performance as I would other sorts of periodic signal to explore the contributing factors.
I thought that this is where Google Trends/insights for search would go. Look at this search: http://www.google.com/insights/search/#q=toys%2C%20mattel%2C... You can see a big spike in searches for 'toys' at the end of every year, which is obviously related to Christmas. You would think that with so much lovely cyclical data, the obvious thing to do would be treat it as a signal, but no.
At first blush, events and trends seem like two different and unrelated things, but it's a fact that almost the entire human race uses the same calendar and clock system, so that both one-off and recurring temporal data both involve fairly straightforward cardinality problems. IT infuriates me that I can search out particular word combinations from billions of documents and get results within a few seconds, but that trying to locate things in time feels like it has hardly advanced at all in over a decade.
Completely agree with you that there's an opportunity for a time based search engine and maybe some standardization so that events can be read by that search engine.
Simple idea but very powerful—you don't need more. They have to put on some filters and integrate some more streams (ie. meetup.com) and then it's perfect.
Yes, completely agree. But I just use this as an indication of the number of people in search of something as opposed to say that if you create it you'll be #1 ranked for each event. Many people will go directly to ticketmaster or meetup as well - those people are part of the market.
On a broader level, there's a huge disruptive opportunity for a time-based search engine. When things happen is often very, very important, but gathering, collating, and connecting this information is painfully difficult. Take a simple example: Federal Reserve Board meetings. There are of tremendous importance to economic analysts, and worth knowing about for any politically aware person. Here is the calendar page: http://www.federalreserve.gov/monetarypolicy/fomccalendars.h... There isn't even an RSS feed. On a more general level, there isn't any obvious tool for browsing recurring events to look at their periodicity. To stick with the economic context (because it's one that matters rather a lot), consider the quarterly earnings and filings of public companies. As someone who likes DSP, it's perplexing to me that I can't add and subtract periodic data about companies' year-on-year performance as I would other sorts of periodic signal to explore the contributing factors.
I thought that this is where Google Trends/insights for search would go. Look at this search: http://www.google.com/insights/search/#q=toys%2C%20mattel%2C... You can see a big spike in searches for 'toys' at the end of every year, which is obviously related to Christmas. You would think that with so much lovely cyclical data, the obvious thing to do would be treat it as a signal, but no.
At first blush, events and trends seem like two different and unrelated things, but it's a fact that almost the entire human race uses the same calendar and clock system, so that both one-off and recurring temporal data both involve fairly straightforward cardinality problems. IT infuriates me that I can search out particular word combinations from billions of documents and get results within a few seconds, but that trying to locate things in time feels like it has hardly advanced at all in over a decade.