I am a Machine Learning Engineer and I am quite frustrated and fed up of the traffic congestion problems in my city (Pune, India) and want to help in solving this problem.<p>Where should I start?
You need to accept that there are a large number of people who don't want to fix the problem, and will actively try to stop you, and that no matter how clever, effective, cheap and world-changing your solution might be you're going to need to become a marketer and a politician to actually get anyone to take it seriously.
You could set up some ASLR cameras, and create a website showing how much the city would make by doing congestion charging. Local governments always love revenue.
I don’t have any domain expertise but I found this project to be pretty cool <a href="https://www.vox.com/2016/8/4/12342806/barcelona-superblocks" rel="nofollow">https://www.vox.com/2016/8/4/12342806/barcelona-superblocks</a><p>In general I guess you’d have to get into public advocacy / local politics space.
Since I am a technologist and not a town planner or a govt. authority I want to propose a solution from technological perspective.
And also I just want to start small by say using web cams to alert/predict future traffic jams and just alert authorities via social media?<p>Then next step would be to feed this systems output to traffic signals and dynamically increase/decrease their timings?
I have a lot of radical ideas and if you have one I would be glad to hear it out.<p>Again keeping in mind that traffic jam eradication vs aversion are a different problems.
Traffic jam eradication/prevention requires a lot of experts from different domain to work in collaboration (town planner + civil engineers + road engineers + gov. authoritarians etc).<p>While I think(may be wrong) traffic aversion can be done using technology + some domain knowledge.
Start by realizing this will require lots of domain knowledge and expertise, machine learning by itself won't solve anything. You will need to work with a team of city planners, civil engineers, etc.