🧪 Open Source AI Project for Soil Analysis – Need Advice and Insights!

AI Discussions
,
1

Hey everyone! I’m a computer engineering undergrad working on an open source initiative to support a laboratory at my university that analyzes soil samples — kinda like a blood test for the earth :seedling:

Their current workflow relies heavily on Excel spreadsheets and manual processing, so I’m building a Python-based system to:

  1. Automate the generation of readable soil reports for farmers (using equations and thresholds)
  2. Apply AI for smart recommendations like fertilizer needs or soil correction (future goal)

Right now, I’m still a beginner in the AI world — I’ve learned Perceptrons, Adaline, MLPs, and I’m hungry to go deeper. Any thoughts on:

  • ML models that could fit this use case?
  • How to approach feature engineering with soil data?
  • Cool tools or examples of similar AI-for-agriculture projects?

This is 100% open-source and intended to help a real-world lab team. I’d love to hear your ideas, questions, or anything you think could help me shape this better :folded_hands:

I’m starting fresh and here is my github repository for this initiative.

Thanks in advance!

2

Deep learning models and Tree models can help with tabular data.

There is a beginners course in Kaggle I came up accross some years ago about feature engineering.

1 Like
3

That sounds like the problem for which Ronald A. Fisher developed statistics.

Not much, and maybe you did: but it always pays to check through the online libraries, your laboratory might have a subscription. Here is one:

1 Like
4

Thanks for the suggestion, I’ll check them out.

5

Yeah the problem sounds like Ronald A. Fisher’s approach. I’ll search through your link. Thanks for the reply!