Sample Project Ideas
All projects require building something with AI/ML. Here's what you could work on:
Build a classifier that detects human emotions from plant bioelectric responses. Train models on voltage time-series data. Compare different ML approaches (neural networks vs. XGBoost vs. random forests).
Tech: Python, TensorFlow/PyTorch, time-series analysis
Create a real-time visualization system that shows plant "mood" as color, sound, or animation. Design algorithms to translate voltage patterns into human-perceivable feedback.
Tech: Python, React/JavaScript, real-time data streaming, sonification
Analyze how bird songs differ between urban and rural areas using existing bioacoustic databases. Build ML classifiers to identify regional "dialects."
Tech: Python, audio processing (librosa), spectrograms, CNNs
Can people learn to "read" plants? Design an experiment where humans listen to sonified plant data and try to identify plant states. Build the sonification system and analyze results.
Tech: Python, audio synthesis, experimental design, statistical analysis
Plants have daily cycles that affect all their signals. Build models that separate circadian patterns from responses to stimuli. Essential for accurate detection systems.
Tech: Python, time-series decomposition, signal processing, Fourier analysis
Build a tool that analyzes your own team's WhatsApp or Slack messages to identify collaboration patterns. Who's a Bee? Who's an Ant? Is anyone slipping into Leech behavior? Create a dashboard that helps teams find groupflow.
Tech: Python, NLP, sentiment analysis, network analysis, visualization
Survey the research landscape of cross-species communication (plants, fungi, bacteria), then build a small proof-of-concept that demonstrates one communication pathway.
Tech: Literature analysis, Python prototype, data visualization
A combination of low-latency software (e.g., SoundJack) and videoconferencing software (e.g., Zoom) enables choir members to sing together online in real time. However, in contrast to physical choir rehearsals, coordination of the group of singers cannot occur through visual signals from a conductor, as the video from conventional videoconferencing software lags behind the audio transmitted by the low-latency software.
This project aims to determine how, if at all, singers can coordinate in the absence of visual cues and measure their level of entanglement.
Tech: Python, analysis of audio and video material, qualitative interviews (if applicable)
Moltbook is an internet forum designed exclusively for AI agents that was recently launched. Similar to Reddit, it allows AI agents to post to the forum and interact with each other, while humans can only observe.
This project uses social media analytics to gain initial insights into Moltbook's structure and interaction patterns. Based on an analysis of features typically found in human-oriented online communities, the goal is to determine how, and if, interactions between AI agents differ from interactions between humans on other social media platforms.
Tech: Python, literature analysis, network analysis, visualization
In the course of the transition to remote work during the COVID-19 pandemic, images of ‘new’ colleagues, including images of pets, were frequently shared on various social media platforms. This project aims to collect and analyse images of pets in the home office and the associated social media posts (e.g. which animals, situations, emotions).
Tech: Python, visual analytics, NLP