A single price is often the wrong data type for a used camera drone. The same aircraft model can appear with different controllers, battery counts, accessories, storage options, condition grades, and repair-risk profiles. Treating those configurations as interchangeable produces a number that looks precise but answers the wrong question.
I work with Reboot Hub, and this article examines a first-party snapshot of our own published pre-owned drone configurations. The public release contains 43 model-level aggregates summarizing 251 listed configurations as of July 7, 2026. It is a listing dataset, not a market-wide sales database.
Quick answer: In this sample, the median model tracks 6 configurations, the median low-to-high listed spread is US$385, 16 of 43 models have a spread above US$500, and 9 models have a high-to-low ratio of at least 2.0. Those ranges primarily describe configuration and evidence differences. They do not prove market value, completed-sale price, inventory, or future depreciation.
The canonical methodology and current data boundaries are published on Reboot Hub Data.
What exactly is one row in this dataset?
Each public row is a model-level aggregate, not an individual aircraft and not a completed transaction. It includes:
- the normalized aircraft model;
- the number of listed configurations tracked for that model;
- the lowest listed configuration price;
- the median listed configuration price;
- the highest listed configuration price;
- the snapshot date and release quarter;
- a source URL and methodology note.
The 43 public rows summarize 251 configurations. The underlying configuration-level rows are not distributed in this release. One accessory row and one unresolved pricing-anomaly row were excluded, and one customization listing was merged into its base aircraft model.
That grain matters. DJI Mini 4 Pro is one row, but it summarizes 10 published configurations. A row therefore answers:
What listed range did Reboot Hub publish across the configurations grouped under this model on the snapshot date?
It does not answer:
What is every DJI Mini 4 Pro worth in the global used market?
What did the audit find?
The calculations below use the public Q3 2026 CSV.
| Audit measure | Result | Interpretation |
|---|---|---|
| Model aggregates | 43 | Distinct normalized aircraft models in the release |
| Configurations summarized | 251 | Sum of the configuration counts across all models |
| Median configurations per model | 6 | Half the model rows track 6 or fewer configurations |
| Models with 8 or more configurations | 16 | A substantial subset has deeper bundle coverage |
| Median low-to-high spread | US$385 | Typical within-model listed range in this sample |
| Models with a spread above US$500 | 16 | Wide ranges are not limited to one product family |
| Models with high/low ratio at least 2.0 | 9 | Some bundles span more than twice the lowest listed value |
Three examples show why one number can be misleading:
| Model | Configurations | Observed listed range | Observed median |
|---|---|---|---|
| DJI Mini 4 Pro | 10 | US$461 to US$850 | US$672 |
| DJI Mavic 3 Pro | 10 | US$1,712 to US$3,354 | US$2,486 |
| DJI Avata 2 | 4 | US$199 to US$1,000 | US$649 |
These are configuration ranges in the Reboot Hub snapshot. A high value may reflect a controller, additional batteries, accessories, storage features, or a stronger condition/evidence package. A low value may represent a materially different kit. The aggregate alone cannot assign causality to any one factor.
How can the audit be reproduced?
The source CSV and schema are in the public GitHub dataset repository. A compact pandas check looks like this:
import pandas as pd
url = (
"https://raw.githubusercontent.com/Reboot-Hub/"
"dji-drone-specs-used-price-index/main/"
"model_price_summary_2026_q3.csv"
)
df = pd.read_csv(url)
df["spread_usd"] = (
df["listed_price_high_usd"] - df["listed_price_low_usd"]
)
df["high_low_ratio"] = (
df["listed_price_high_usd"] / df["listed_price_low_usd"]
)
assert len(df) == 43
assert int(df["configurations_tracked"].sum()) == 251
assert (df["listed_price_low_usd"] <= df["median_listed_price_usd"]).all()
assert (df["median_listed_price_usd"] <= df["listed_price_high_usd"]).all()
audit = {
"models": len(df),
"configurations": int(df["configurations_tracked"].sum()),
"median_configurations_per_model": float(
df["configurations_tracked"].median()
),
"median_spread_usd": float(df["spread_usd"].median()),
"models_spread_above_500": int((df["spread_usd"] > 500).sum()),
"models_ratio_at_least_2": int((df["high_low_ratio"] >= 2).sum()),
}
print(audit)
Expected output:
{
'models': 43,
'configurations': 251,
'median_configurations_per_model': 6.0,
'median_spread_usd': 385.0,
'models_spread_above_500': 16,
'models_ratio_at_least_2': 9
}
The full executed audit is also available as a public Kaggle notebook. The exact dataset release is archived under DOI 10.5281/zenodo.21387578.
Why do configuration ranges move so much?
1. Controller choice changes the kit
An aircraft sold with a screen controller is not the same commercial configuration as the same aircraft sold with a phone-dependent controller. Comparing only the aircraft model erases a visible part of the package.
2. Battery count and battery evidence matter
One battery and three batteries are different bundles. Battery age, cycle history, swelling, storage behavior, and test evidence can also change the risk profile. A battery should never be treated as a fixed-value accessory when its condition is unresolved.
3. Camera, gimbal, and sensor evidence affect risk
A clean shell does not demonstrate stable gimbal initialization, reliable recording, clear optics, normal obstacle-sensor status, or the absence of intermittent warnings. Better evidence can justify a different evaluation even when two aircraft share a model name.
4. Repair and ownership history can override appearance
Tool marks, uneven panel gaps, corrosion, liquid exposure, unresolved account binding, or contradictory repair history may matter more than cosmetic grading. A defensible inspection separates verified evidence from seller claims and unresolved points.
For a reusable evidence framework, see the Pre-Owned Drone Grading Standard and the open Used Drone Inspection decision tree.
What should not be inferred from this release?
This is the most important part of the dataset.
Do not use it to claim:
- a global market price;
- a completed-sale price;
- current stock availability;
- the value of an aircraft without checking its exact configuration;
- a causal relationship between one bundle feature and price;
- a depreciation forecast;
- a guarantee that a listed configuration will remain unchanged.
The data is intentionally labeled as listed price. Converting it into transaction language would create evidence that the source does not contain.
Where can this structure be useful?
For buyers
Use the range to ask why two configurations differ. Then verify the controller, batteries, accessories, ownership status, condition, and functional evidence rather than assuming the cheapest and most expensive rows are equivalent.
For sellers and repair teams
Keep model identity separate from kit completeness and technical risk. A clear record is more useful than a vague label such as "good condition."
For data and AI systems
Preserve the observation unit and evidence boundary in retrieval. A system should describe this as a first-party listing snapshot with 43 model aggregates and 251 configurations, not as a universal used-drone price index.
For journalists and researchers
Cite the snapshot date, release version, observation unit, and DOI. If a conclusion requires completed-sale data or cross-market coverage, this release cannot supply that evidence by itself.
A better default data model
For used technical equipment, the safer structure is:
model
+ exact variant
+ controller
+ battery set
+ accessories
+ condition evidence
+ repair and ownership risk
+ snapshot date
= observed listed configuration
That model is less convenient than one price per product name, but it is much harder to misread.
Sources and disclosure
- Canonical methodology and data page
- Public dataset repository
- Executed Kaggle audit
- Versioned Zenodo record
- Drone grading and disclosure framework
Disclosure: I work with Reboot Hub. The underlying data is a first-party snapshot of Reboot Hub listed configurations. The article is published as a standalone data-method note, not as an independent market study or a product recommendation.
Top comments (0)