Ask HN: I want to solve traffic congestion problem, where should I start?
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.
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 https://www.vox.com/2016/8/4/12342806/barcelona-superblocks In general I guess you’d have to get into public advocacy / local politics space. This is amazing, but as a technical guy I cannot change/plan/design city routes (I wish I could!). I forgot to add to the question description but I have added in the comment, that I cannot eradicate traffic jams as that involves government/politicians/city planner etc, while traffic aversion is more dependent on using technology, I thinks that's where I fit. But anyways thanks for link, I got to know something that I had not known earlier. 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? 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. 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). 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. I never said I want to use machine learning (if it needs to be used I am well equipped to use it)! As for domain knowledge I think that can be learned with time (this is a long term project), I have worked on three domains (namely, Advertising, Retail and Finance) without having any clue about it at first. So what I want is theories/books on this problem and how other people have solved it successfully/ not successfully etc...
Ofcourse 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?