Understanding Equipment Data

Mixed-fleet data is inherently complex. Integrating directly with multiple OEMs means navigating a constantly shifting landscape of APIs, data formats, and quality issues.

This guide explains the challenges Hiboo solves on your behalf and provides best practices for building robust applications on the unified data we provide.

Challenge 1: Availability & Timing

Getting timely data from every machine is the first major hurdle. Availability is affected by OEM systems, machine state, and connectivity.

How Hiboo Helps

We manage a complex orchestration of API calls, respecting each source's rate limits and patterns to retrieve data as efficiently as possible. We normalize timestamps to UTC and handle out-of-order data to present a coherent timeline.

OEM API Evolution Management: In 2023-2024, we monitored approximately 20 notable changes each year across major OEM APIs. We absorb breaking changes, deprecated endpoints, authentication updates, and data format modifications so your integration with the Hiboo API remains stable.

What You Should Know

Even with our optimizations, the underlying source determines the data's cadence:

Varying Frequency: Some sources provide multiple data points per minute, some provide only one data point per day. Your application should not a predefined data frequency for every machine.

Varying Freshness: Some sources provide near real-time updates, while others send data in hourly or daily batches. Your application should not assume a constant, real-time stream for every machine.

Connectivity: Equipment in remote locations will inevitably have delayed data.

Engine-Dependent Communication: Many machines may only transmit data when the engine is running. This means you should expect data gaps during off-periods (nights, weekends). Data generated just before engine-off might also be delayed until the next time the machine is on.

Time Zone Variations: OEMs may report in UTC, local time, or equipment's current location time with different formats (ISO 8601, epoch, custom formats).

Challenge 2: Quality & Accuracy

Once data arrives, its quality can vary dramatically. Raw data from the field is often noisy and contains errors.

How Hiboo Helps

Our data pipeline includes a robust validation and cleaning layer. We automatically detect and, where possible, correct common issues like:

Invalid GPS Coordinates: Filtering out erroneous locations (e.g., 0,0).

{
  "latitude": 0,
  "longitude": 0,
  "timestamp": "2024-01-15T10:30:00Z"
}

Decreasing Counters: Identifying and flagging impossible values like decreasing odometers or hour meters.

{
  "cumulativeOperatingHours": 1250.5,  // Previous reading
  "cumulativeOperatingHours": 1248.2   // Current reading (impossible)
}

Sensor Errors: Detecting values that are physically impossible (e.g., fuel levels over 100%, extreme temperatures).

We provide a Signal Health Dashboard in our platform to give you visibility into the quality of the underlying data streams.

What You Should Know

While we clean obvious errors, some ambiguity is inherent. A sensor might be poorly calibrated but still produce data within a "valid" range. Your system should treat the data as a highly reliable indicator, but be prepared for occasional, unfixable anomalies from the source.

Data Duplication: Same data may be sent multiple times after connectivity loss - we handle deduplication.

Partial Updates: Equipment systems don't always send complete data sets. We merge partial updates without losing previous values.

Challenge 3: Consistency & Standardization

Each OEM has its own proprietary data format, field names, and units, making cross-fleet analysis nearly impossible without a normalization layer.

How Hiboo Helps

This is our core value. We map dozens of disparate sources to a single, unified data model based on the ISO 15143-3 (AEMP) standard.

Unified Naming: engineHours vs. operatingHours all become cumulativeOperatingHours.

Standardized Units: We convert everything to metric standards (hours, liters, Celsius).

API Stability: We absorb the ongoing API changes from OEMs. If a source changes its authentication, deprecates an endpoint, or alters a data format, we handle it. Your integration with the Hiboo API remains stable.

Beyond the Standard: We extend the AEMP standard to support data types it doesn't cover, such as for electric vehicles, light vehicles, and power generators.

What You Should Know

Not all data kinds are available for every machine. A 10-year-old excavator will not provide the same rich data as a new electric one. When querying data, always check for the presence of a field before using it.

• Subscription Dependencies: Higher OEM subscription tiers often include more data types.

• Model Year Differences: Newer equipment typically provides richer data.

Best Practices for Building on Hiboo

Building with Hiboo means you're already 95% of the way there. To build truly resilient applications, embrace the nature of real-world equipment data:

Handle Nulls Gracefully: Your code should never assume a data field will be present. Always check for null or undefined values.

Focus on Trends, Not Just Points: A single data point might be an anomaly, but a trend over time is almost always reliable. Base critical logic on rolling averages or time-series analysis.

Use Realistic Time Windows: When querying for data, don't assume everything arrived in the last 5 minutes. Use wider time windows (e.g., the last 24 hours) to account for batching and delays.

Trust, But Verify: For critical workflows (like automated billing based on usage), consider adding a "reasonableness" check in your own logic as a final safeguard.

Consider Business Context: Align expectations with operational realities - perfect data is rare, but data is often good enough for operational decisions.

Understanding these realities helps build more robust, realistic systems that work effectively with real-world mixed fleet data.