Give AI the graph image
Temperature, relative humidity, air VPD, PPFD, the numerical summary, and daily DLI are combined into one image.
ENVIRONMENT DATA FOR AI DIALOGUE
Choose a day or period, copy the graph and prompt with one click, then paste them into the AI you already use.
Add the plant name and anything you have noticed, and use the measurements as a starting point for the conversation.
| Paste step 2, then step 3, into the same AI conversation.
Temperature, relative humidity, air VPD, PPFD, the numerical summary, and daily DLI are combined into one image.
A question prompt will be prepared for the “Think through it together” style.
If your AI tool cannot accept a directly copied image, save the graph image and attach the file instead.
Temperature, relative humidity, air VPD, and PPFD are aligned to the same JST timeline.
Kyoto Horticultural Base · Environmental data for AI dialogue
Temperature °C | Relative humidity %
kPa | Calculated from display-processed temperature and relative humidity
µmol m⁻² s⁻¹
Calculated from valid values after browser-display aggregation.
| Measure | Minimum | Maximum | Mean | Unit |
|---|
Integrated from one-minute PPFD values between 04:00 and 20:00.
Temperature, relative humidity, and PPFD are read from the same public JSON data derived from the source measurements used elsewhere on Kyoto Horticultural Base. Air VPD is calculated from the display-processed temperature and relative humidity. DLI is integrated from one-minute PPFD values between 04:00 and 20:00.
Temperature, relative humidity, and PPFD are range-checked; invalid values are treated as missing. Long gaps are not connected. PPFD gap checks and DLI integration use 04:00–20:00, and unmeasured values are not assumed to be zero.
Only short gaps of up to five minutes are linearly interpolated. Temperature and relative humidity use a five-minute moving average; PPFD is not smoothed. DLI is calculated from one-minute PPFD after short-gap interpolation. The source public JSON is not modified.
If multiple valid values exist for the same measure at the same timestamp, their mean is used for display processing. Duplicate counts are retained with the quality information.
These graphs show air and light measured at one observation point in Fushimi, Kyoto. They are not measurements from the exact place where your plant is growing. Leaf temperature, pot temperature, root-zone moisture, wind, pests and disease, substrate, cultivar differences, and acclimation after purchase are not measured here and may need to be added as local context.
USING DATA WITH AI
Once you give AI the graph image and question prompt, changes in temperature, humidity, and light can be considered alongside the plant’s day. The graph shows only part of the environment around a plant. Adding the plant name, where it is growing, and anything you have noticed can make the conversation much more specific.
You do not need to prepare every detail in advance. Start with a short question. After you paste the graph image and prompt into AI, add information as needed: the plant name, symptoms, when they appeared, where the plant is kept, sun exposure, shade, airflow, whether it is in a pot or in the ground, soil or potting mix, watering, fertilizer, and when the plant was purchased. If you do not know something, do not guess. AI can ask follow-up questions and you can work through the situation one step at a time.
These graphs come from one observation point in Fushimi, Kyoto. Even within Kyoto, conditions change with buildings, pavement, direction, elevation, wind, and shade. A balcony, garden, indoor room, or sheltered spot can feel very different. Rather than treating these measurements as the exact conditions around your plant, tell AI what seems similar and what seems different. Those differences can be useful clues.
The graphs show air temperature, relative humidity, air VPD, and light at the observation point. They do not directly show leaf temperature, temperature or moisture inside the pot, root condition, wind, pests or disease, soil characteristics, fertilizer concentration, cultivar differences, conditions during transport and retail, or how well a recently purchased plant has acclimated. A change in the graph and a plant symptom happening at the same time does not prove that one caused the other.
Use AI’s answer as a clue for further observation rather than as a final conclusion. Ask whether the proposed explanation fits what you actually see, what other possibilities remain, and what you could observe next to distinguish between them. For pesticides, human or animal safety, or expensive treatments, confirm important decisions with product labels, specialist guidance, or primary sources as well.
Instead of asking only “What is the cause?”, try asking: “Please separate what can be read from this graph, what cannot be determined from it, possible explanations, and what additional information would help.” When thinking about future conditions, do not treat past measurements as a forecast. A better question is: “If similar conditions occur again, what effects might be possible?”
The measured data sent to AI is the same whichever style you choose. What changes is the way the conversation is guided.
Starts with the most important features of the graph in plain language. Technical terms are explained briefly when needed. If important information is missing, AI asks for one particularly useful detail. This is a good choice if environmental data is new to you or you mainly want the key points first.
Works more like a gardening conversation. AI begins with what can be seen in the graph and several possible explanations, then asks about the plant, its location, sun exposure, watering, soil, airflow, purchase timing, and other relevant details. The next question changes according to your answer, so the goal is to understand the situation step by step rather than rush to one cause.
Starts by checking data availability and missing-data status, then examines temperature, humidity, air VPD, PPFD, and DLI in more detail. It also distinguishes sudden changes, missing values, possible sensor issues, measured facts, scientific explanation, conditional inference, and things that are still unknown. This is useful when you want to inspect the evidence and limitations carefully.
Does not impose a special response length or order. You continue with the AI you normally use. If that AI remembers earlier conversations, it may be able to use information you have already shared about your plants and growing conditions. This is useful when you want to bring the measured data into an ongoing conversation.