SUPER Beginner Here - How Do I Start Making a Simple Sales Route Mapping App?

Hi everyone,

I’m a big fan of all the incredible tech and tools I’ve come across here, but I’ll admit, I’m a total beginner when it comes to building anything myself. That said, I’m really excited about the idea of creating a simple app to help me efficiently plan my daily sales routes.

I picture something that lets me input a list of addresses, and the app would map out the most efficient driving route to save time and gas. Nothing too fancy—just functional and easy to use.

Here are a few questions I have:

  1. What’s the best way for someone with very little coding experience to start a project like this?
  2. Are there any tools, frameworks, or tutorials you’d recommend?
  3. Would something like Google Maps API be a good starting point for this?
  4. Am I biting off more than I can chew as a beginner, or is this a reasonable project to tackle?

Thanks in advance for any advice, resources, or words of encouragement. I’m eager to learn and start tinkering!

Thanks
Sonny

1 Like

This is known as the travelling salesperson problem. If you Google finding approximate solutions using ML, that might be a start.

However, in general, this is an unsolvable problem (ie it’s exponential cost in the number of possible route, but there are efficient approximate solutions).

2 Likes

Here’s an example : Solving the Traveling Salesperson Problem with deep reinforcement learning on Amazon SageMaker | AWS Open Source Blog

2 Likes

and here’s a research survey: The neural network methods for solving Traveling Salesman Problem - ScienceDirect

2 Likes

Wow, this is intense—definitely a bit above my skill level, but I get the gist. My naivety is showing, guess I’ll keep at it as usual. On the bright side, at least I’m not relying on a Thomas Guide like when I first started traveling!

Thanks so much for taking the time to answer my question.

1 Like

Just because the problem is theoretically intractable, it doesn’t prevent the approximate solutions from being good - or that using it as a test case to learn how to do machine learning is diminished. I definitely found it an interesting question. But yes, if you want something slightly simpler to cut your teeth on, then something that is more easily formulated as a Yes/No problem, and ideally can leverage pre-existing data sources for ground truth will be easier (think cat or not cat :)).

1 Like