Analytics & ML · Study deck
Edge Data: Sampling and Statistical Compression
This first route protects signal evidence through sampling and compares lossless, statistical, and transform reduction.
Data Dora is your guide for this deck.
After studying this chapter
Learning objectives
You will be able to:
- Apply Nyquist Theorem: Calculate appropriate sampling rates for different sensor types
- Implement Data Reduction Techniques: Use aggregation, compression, event-based reporting, and delta encoding
- Select Compression Algorithms: Choose optimal algorithms based on data type and edge device constraints
- Avoid Common Pitfalls: Prevent sampling aliasing, buffer overflow, and rate mismatch errors
Major section
Start With the Story
If the device measures too rarely, no later data trick can rebuild the missing motion.
- Sampling means taking readings at chosen times.
- The sampling rate is how many readings are taken in one second.
- Compression means using fewer bits to carry useful information.
- Exact records may need a lossless method, which can rebuild every bit.
Major section
Edge Sampling Compression Basics
Edge sampling and compression reduce the amount of data IoT devices need to transmit.
- By transmitting only important changes or compressed summaries, devices save battery power and network bandwidth while preserving the information that matters most.
Major section
Edge Data Reduction Basics
Core Concept: Transform raw sensor data into actionable information at the source - send summaries, statistics, and alerts rather than every reading.
- Byte and radio-energy savings depend on payload and protocol overhead.
- An average loses within-minute peaks and timing, so retain other features or raw windows when the application needs them.
Major section
Key Concepts
Adaptive sampling: Dynamically adjusting the sensor sampling rate based on signal variance or event rate — increasing frequency when the signal changes rapidly and decreasing it during quiet periods.
- Delta encoding: A compression technique transmitting only the change between consecutive readings rather than absolute values, highly effective for slowly varying sensors.
Major section
Phoebe's Field Notes: Why Sampling First Protects Compression
For this exercise, suppose a hypothetical source produces eight impulses per shaft revolution: its repetition rate is 240 Hz, with harmonics at 480 Hz and 720 Hz.
- Strict Nyquist for the third harmonic needs 1,440 Hz, and the chapter's 2.5× margin needs 3,600 Hz.
Major section
Putting Numbers to It
Illustrative impulse frequency (not a bearing-defect prediction):: If a source produces eight impulses per 30 Hz shaft revolution, its repetition rate is 8×30=240 Hz.
- Nyquist frequency = 50 Hz, but the illustrative component is at 240 Hz.
Major section
Try It: Edge Aggregation Data Reduction
Adjust the sensor sampling rate and aggregation window size to see how edge aggregation reduces data volume.
- The widget calculates the bandwidth reduction and shows what information is preserved versus lost.
Deck summary
Key takeaways
If the device measures too rarely, no later data trick can rebuild the missing motion.
- Edge sampling and compression reduce the amount of data IoT devices need to transmit.
- Core Concept: Transform raw sensor data into actionable information at the source - send summaries, statistics, and alerts rather than every reading.
- Adaptive sampling: Dynamically adjusting the sensor sampling rate based on signal variance or event rate — increasing frequency when the signal changes rapidly and decreasing it during quiet periods.
- For this exercise, suppose a hypothetical source produces eight impulses per shaft revolution: its repetition rate is 240 Hz, with harmonics at 480 Hz and 720 Hz.
Retrieval practice
Recall check 1 of 2
Data Dora says: answer from memory, then check your reasoning.
Q1A 60 Hz shaft signal has illustrative components at the 2x, 3x, and 4x harmonics. What is the nominal Nyquist rate (twice the highest frequency of interest) before adding a practical sampling margin?
Show answer
Answer: D The nominal Nyquist rate is twice the highest frequency of interest: 2 × (4 × 60 Hz) = 480 Hz.
Retrieval practice
Recall check 2 of 2
Data Dora says: answer from memory, then check your reasoning.
Q2A binary contact stays closed for many records. The receiver must restore the exact state sequence. Which reduction fits?
Show answer
Answer: A Repeated equal states can be represented without losing the sequence.
Print reference
Answers
Answer key.
- D · The nominal Nyquist rate is twice the highest frequency of interest: 2 × (4 × 60 Hz) = 480 Hz.
- A · Repeated equal states can be represented without losing the sequence.