![]() It lets you answer questions like “how big is Greenland, really?” or “how large would Alaska be if it were in the contiguous US?”Ĭomparea projects the two geographic features using equal area projections with the same scale but different centers. You can read more on Comparea’s about page or study this nifty diagram: This results in a valid comparison of their areas with minimal distortion for each shape. It works great on both desktop and touch devices, where you can drag the two shapes with your fingers or pinch to zoom. Like most of my side projects, this one was developed on and off over a long period of time. I first worked on the project in spring 2012, discovered Natural Earth Data and quickly made a purely client-side demo. The response was generally positive, but I didn’t release it because of general UI jankiness (the transitions never quite worked), the lack of a backend and the feeling that I didn’t have enough polygons to make it fun.Ībout a year later (in early 2013) I started playing with Open Street Maps data. There are tons of great polygons in their data dumps, including smaller features like Golden Gate Park or Central Park. I spent a few weeks playing around with this data set, but ultimately could’t come up with a good way of deciding which shapes were notable enough to include (there are millions, and not every one block park is notable enough for inclusion). I’d shown the Comparea demo to enough people that I was convinced it was worth publishing. (One notable demo session was with my six year old nephew, who rattled off country comparisons to me for nearly an hour!) When I got some time off between jobs, I decided to make releasing Comparea an explicit goal. In addition to the finished product, the real value of any side project is what you learn from doing it. Flask This was my third Flask project (after gitcritic and webdiff) and I finally knew enough to organize it correctly.This was my first Heroku project and my first time using CloudFlare for distribution. Developing using Flask and Heroku is far more lightweight and flexible than using AppEngine, which was my tool of choice in the past. Its projection and behavior tools were godsends. These both made lots of gross SVG manipulation code from my initial demo melt away. I debugged Natural Earth Data shapes using IPython Notebooks.
0 Comments
Leave a Reply. |