# Soil Properties Prediction using Vis-NIR spectra and Machine Learning Our RapidAPI project offers a solution for determining soil properties using machine learning models. This implementation is based on Vis-NIR reflectance spectra data and has been trained on a dataset of 448 wavelengths between 400nm and 1000nm, along with the corresponding soil properties. The use of various machine learning algorithms ensures high accuracy in predicting 74 different soil properties. Integration with the RapidAPI platform is seamless, allowing clients to send a post request with the Vis-NIR reflectance spectra data in JSON format and receive the predicted soil properties values in real-time. This eliminates the need for costly and time-consuming traditional lab analysis, which can range from $20 to $200 per sample and take several days to weeks to receive results. By using our API, users can quickly and easily obtain results for 74 different soil properties by simply uploading a Vis-NIR spectra file, which can be obtained in minutes using a spectrometer. The API provides results in real-time, saving both time and money. Additionally, the API's predictive models have an R2 score ranging between 80% and 99%, ensuring high accuracy and precision in the predicted soil properties. In conclusion, our RapidAPI implementation is a cost-effective, fast and accurate solution for determining various soil properties. It provides real-time results and eliminates the need for traditional lab analysis, saving both time and money for users. Link to a text document containing a json exemple of a reflectance spectral signature in the desired format : https://drive.google.com/file/d/1dj5zxhyPWE5iIct12dscfujSSun_d_3f/view?usp=sharing Link to a text document containing the 74 available soil properties for prediction : https://drive.google.com/file/d/1SB-pZwjHppcGFVs2ppSCCauXTpK_c5rH/view?usp=sharing