Bridging IoT and Machine Learning with Python for Sustainable Agriculture
summary
An IoT sensor network with microcontroller and API collects soil pH NPK temperature and humidity plus open‑weather data over a 75-day field trial, producing a locally sourced dataset of 189,729 rows and 58 columns. Data cleaning uses mean imputation and numeric encoding, feature selection via SelectKBest mutual information and RF selector, and models include XGBoost MLP KNN and bagging. XGBoost with selected features reaches near‑perfect fit (R2 ≈ 1) and the system reports a claimed 35% yield increase.
AI-written summary · about
chapters (10)
- 0:00IoT and ML in agriculture
- 4:27IoT sensors collect farm data
- 9:49Python in smart farming
- 13:21Maize yield prediction case
- 21:44Training yield prediction models
- 24:39Results and smart farming plans
- 25:22Q&A: Data correlation and evaluation
- 26:45Public data and region expansion
- 28:10Q&A: Sensor selection and lifetime
- 30:37Closing remarks
related talks
28mHarnessing Machine Learning to Improve Agricultural Resilience in Africa
10mUsing Machine Learning to Reduce World Hunger
26mAI, big data, digital and future of agriculture
28m"from seeds import plants: using IoT to grow healthy herbs"
19mPlant Factory: Sensor, Data, Machine Learning
07mAI Infrastructure Transforming Modern Agriculture
25mFrom seed to store: AI to optimize the indoor farms of the future
17mJeff Bradshaw, Founder, Adaptris. NYC 2017