Catapolt.co: A web app for making awesome polls
I am an avid sports fan and I probably have participated in thousands of polls in my lifetime asking which player is the best or what my predictions are for the season, etc. The one thing all these polls share in common is they are really boring and not very sophisticated. I wanted to create a site that addresses both of those issues. Not really out to build a business, although that is an aspirational goal at this point. Of course, a site like this needs plenty of users to make sense.
The site is called Catapolt (yeah, it's a silly but hopefully memorable pun on cats, polls, and catapults lol). The frontend is built with VueJS and Vuetify material design UI frameworks. The backend is django/django-rest-framework. (Awesome combo so far.)
The idea of the app is to create and share really large polls (i.e. with many, perhaps, even dozens or hundreds of options). The voting style is "Choose A-or-B" and I use Elo ratings (as used in chess) to calculate ratings, rankings, and win probabilities.
Eventually, I would like to incorporate more "AI" to automate the creation of polls, which right now is a manual (and admittedly painful) chore. But participating in a poll is a lot of fun I'm told.
Would appreciate suggestions on any design elements or features, and if have suggestions for how to market such a site! Also feel free to ask questions about the site and how I built it.
Congrats on launching! One question on your value proposition:
Can you elaborate on why sophisticated polls are valuable? To me, normal polls get the job done in a clear straightforward manner. I tried the preview here and was immediately confused about the win percentage.
Oh, about the win probability. Given the two ratings (which I hide until the poll is over), the win probability essentially tells you the likelihood that the option you selected would have won that matchup based on the current (Elo) ratings. So, for example, if WP=80%, that means you picked the option that had an 80% chance of winning that matchup (i.e. was much more popular).
Yeah, think about a poll for drafting players in a sport (as you can see there is an NBA Draft poll on the site). Typically the polls I see will list 4 or 5 players who everyone assumes could be the #1 pick. So the poll will be super simple and focused on that one slot.
But in reality the Draft is interesting from spots 1-60 and fans care about each ranking, especially as it relates to their favorite team. So my site gives you a way to rank dozens of players and get real information about where they rank relative to each other. Not just the top few slots (which you could get by doing a bunch of different polls).
You could say, well, put all 60 players in one giant radio button poll, but like 50 of those players would get 0 votes for the #1 pick. So you'd get zero information on their relative rankings.
Clearly, this same logic applies to other things like movies, books, etc. When you really care not just about the #1 thing, but all the options.
(Also, FWIW, ESPN periodically puts out their #NBARank surveys of the league's top 100 players and having been a member of the voting panel, this "Choose A-or-B" style of poll is how they did it. They probably use Elo too. But it's just a very boring text-based poll.)