In a compelling presentation at MJBizCon 2024, Fluence Data Scientist Dr. Jason Matlock addressed one of the cannabis industry’s most pressing challenges: how to effectively use data to improve cultivation operations without becoming overwhelmed by complexity. As the cannabis industry continues to commodify and profit margins tighten, the ability to make data-driven decisions has become increasingly crucial for success.
“The key here is to recognize that collecting data is not the same as gaining insight,” Matlock emphasizes. “There’s a lot of companies who are collecting data, and they’re drowning in it. They have all sorts of records, but they’re not learning anything from that material.”
Verständnis der Datenwissenschaft im Cannabisanbau
Matlock räumt mit dem weit verbreiteten Missverständnis auf, dass Datenwissenschaft komplexe KI-Systeme oder neuronale Netze erfordert. Stattdessen stellt er die Datenwissenschaft als einen zugänglichen Prozess dar, der damit beginnt, die richtigen Fragen zu stellen. Einfache Visualisierungstools wie Boxplots und Liniendiagramme können bei strategischem Einsatz oft aussagekräftige Erkenntnisse liefern.
Arten der Datenanalyse
In der Präsentation werden zwei grundlegende Arten der Datenanalyse im Anbau beschrieben:
- Untersuchung: befasst sich mit bestimmten Ereignissen oder Ergebnissen und versucht zu verstehen, warum etwas passiert ist, um Erfolge zu wiederholen oder Misserfolge zu verhindern
- Prozessüberwachung: Konzentriert sich auf die Sicherstellung der konsistenten Ausführung bekannter Best Practices
Fallstudien zur Analyse von Cannabisdaten
To demonstrate how data science principles can be effectively applied in cannabis cultivation, Dr. Matlock presented two compelling case studies. These examples showcase both investigative analysis and process monitoring in real-world scenarios, illustrating how systematic data collection and analysis can solve complex cultivation challenges. The first case study examines a temperature management issue that challenges common assumptions, while the second explores innovative approaches to standardizing defoliation practices. Both cases highlight the practical value of data-driven decision-making in modern cannabis cultivation.
Fallstudie 1: Untersuchung zum Wärmemanagement
To illustrate these concepts, Matlock shared a practical example involving a client’s claim about Fluence fixtures causing excessive heat in their greenhouse. Through systematic data analysis using simple box plots, the team discovered that room location and insulation, not the fixtures, were responsible for temperature differences. This case study demonstrated how asking the right questions and following a methodical approach can reveal unexpected insights and solutions.
Fallstudie 2: Überwachung des Entlaubungsprozesses
The second example focused on process monitoring in defoliation practices, a critical but labor-intensive cultivation task. “How do we know if defoliation is standardized and under control?” Matlock asks. “If it’s that critical and we’re investing that much into it, why don’t we know objectively if we are successfully standardizing this practice?”
He presented two potential measurement approaches:
- Überwachung des Lichteinfalls in das untere Kronendach
- Automatische Bildgebung zur Zählung sichtbarer Knospenstellen
Die Wahl zwischen diesen Methoden hängt von der Betriebsgröße und den Ressourcen ab, was zeigt, wie Data-Science-Lösungen auf unterschiedliche Anbauumgebungen zugeschnitten werden können.
Prozess der Umsetzung
In der Präsentation wurde ein klarer Prozess für die Implementierung von Data Science in der Landwirtschaft skizziert:
- Beginnen Sie mit spezifischen, geschäftskritischen Fragen
- Anwendung des Systemverständnisses zur Ermittlung relevanter Metriken
- Entwicklung geeigneter Datenerhebungsmethoden
- Daten verarbeiten, um aussagekräftige Muster hervorzuheben
- Analysieren Sie die Ergebnisse und ergreifen Sie Maßnahmen
- Wiederholen Sie den Zyklus, um die Auswirkungen zu messen.
Schlussfolgerung
“Data science at its core is meant to be accessible and empowering,” Matlock concludes. “Start simple, ask intentional questions, take steps to measure and refine what matters the most, and you’ll be successful. You can always build from there, but if you don’t start at that point, you almost certainly will become lost.”
The presentation effectively demonstrated that data science in cannabis cultivation doesn’t require advanced degrees or complex systems. Instead, it demands a methodical approach to asking and answering questions that directly impact business success. By following this systematic process and starting with simple, focused metrics, cultivators of any size can begin leveraging data to improve their operations and bottom line.
This practical approach to data science offers cannabis cultivators a clear path forward in an increasingly competitive market, where efficiency and consistency are paramount to success. The emphasis on accessibility and intentionality provides a refreshing counterpoint to the often-overwhelming world of big data and artificial intelligence, making data-driven decision-making achievable for operations of all sizes.
Über den Autor
Dr. Jason Matlock brings over 10 years of experience in the cannabis industry, including commercial scale cultivation, facility design and construction, and business development consultation. He excels at integrating and balancing economic, logistical, and horticultural considerations when addressing operational challenges. Dr. Matlock has a proven record of success with designing and conducting grower-participatory on-farm research trials, and is skilled at performing statistical analyses on biological datasets and presenting results in easily interpretable formats.
