The first one took way too long due to PHP, my webhost, so-called APIs and general coding rustiness. It quickly became apparent that when working on larger datasets, making a lot of requests to APIs early was just going to make people think things weren't working. It also became apparent that iFrames suck and it was one more headache I didn't need. Thirdly after investing a week of free time into it, I knew I needed to invest more to get a result I could be proud of. This is yet another intermediary step, which uses my first ever PHP class.