Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
15 views

Introduction To Data Analytics For IoT

The document discusses data analytics for IoT and the challenges involved. It describes how IoT devices generate huge amounts of structured and unstructured data. It also covers challenges like handling real-time streaming data, scaling databases to the data volumes, and analyzing unstructured data like images and video.

Uploaded by

onshop.online.in
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
15 views

Introduction To Data Analytics For IoT

The document discusses data analytics for IoT and the challenges involved. It describes how IoT devices generate huge amounts of structured and unstructured data. It also covers challenges like handling real-time streaming data, scaling databases to the data volumes, and analyzing unstructured data like images and video.

Uploaded by

onshop.online.in
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Introduction to Data Analytics for IoT

Contents:
● Introduction (Data Generation in IoT).....................................................................2
● Structured Versus Unstructured Data………………………………………………2
● Data in Motion Versus Data at Rest………………………………………………..3
● Types of Data Analysis in IoT……………………………………………………..3
● IoT Data Analytics Challenges…………………………………………………….4
● Conclusion…………………………………………………………………………4
IoT (22MCA32)

Introduction (Data Generation in IoT)


IoT devices are things like sensors, machines, or gadgets that are connected to the internet. These
devices constantly generate huge amounts of data.
For example, modern jet engines on airplanes have many sensors that produce 10 gigabytes (GB)
of data every second while flying. That's a lot of data! 1 GB is enough to store hundreds of
photos or a few hours of video.A single commercial airplane flying can create over 500 terabytes
(TB) of data in a day from all its sensors and systems. 1 TB is enough to store around 500 hours
of movies.The entire commercial airline industry, with all the planes flying everyday, can
potentially produce petabytes (PB) of IoT data daily. 1 PB is a million gigabytes! That's an
unimaginably massive amount of data.

Similarly, smart utility meters in homes that measure electricity, gas etc. can collectively provide
over 1 billion data points or readings every single day from all the meters across a region or
country.With IoT devices producing such mind-boggling volumes of data constantly, it becomes
very important to analyze this data efficiently using data analytics techniques and tools. This
analysis helps extract useful insights and information from the huge IoT data flow.Without proper
data analytics, all the IoT data being generated would be wasted and provide no real value. Data
analytics helps make sense of the IoT data deluge.

Structured Versus Unstructured Data


Structured data is organized and follows a predefined model or schema, such as data in databases,
spreadsheets, or sensor data. Unstructured data, on the other hand, does not conform to a model,
like text, images, video, or speech. Around 80% of business data is unstructured. IoT devices
generate both structured and unstructured data. Structured data is easier to process using
traditional analytics tools, while unstructured data requires advanced techniques like machine
learning and natural language processing.

Example: A smart home security system generates structured data like sensor readings
(temperature, motion, etc.) but also unstructured data like security camera footage and audio
recordings.

Department of MCA 2
IoT (22MCA32)

Data in Motion Versus Data at Rest


Data in motion is data that is in transit or being transmitted over the network, such as
client/server exchanges or IoT device data streams. Data at rest is data that is stored or persisted,
like in databases, storage arrays, or data centers. In IoT, device data is initially "data in motion"
until it is processed at the edge or forwarded to data centers or the cloud. Data analytics tools
differ for "data in motion" (stream processing) and "data at rest" (batch processing on stored
data).
Example: A connected car generates data in motion (sensor readings, GPS data) that needs
real-time processing. The processed data is then stored as data at rest for further analysis.

Types of Data Analysis in IoT


Descriptive Analysis: Answers "What is happening?" by analyzing raw data and events, such as a
truck engine's temperature.
Diagnostic Analysis: Answers "Why did it happen?" by analyzing the root causes of events or
issues.Predictive Analysis: Aims to forecast future events or issues based on historical data
patterns.Prescriptive Analysis: Provides recommendations on actions to take for predicted events
or issues.
Example: A predictive maintenance system for industrial equipment uses descriptive analysis to
monitor current performance, diagnostic analysis to identify potential issues, predictive analysis
to forecast future failures, and prescriptive analysis to recommend maintenance actions.

IoT Data Analytics Challenges

Traditional Databases Limitations: Scaling issues due to the high volume of IoT data and
inflexible schema unsuitable for evolving IoT data models. The solution is to use NoSQL
databases like MongoDB.
Real-time Streaming Data Analysis: High-velocity IoT data streams require real-time analysis.
Edge analytics (like Azure IoT Edge, AWS Greengrass) is needed to detect patterns and
anomalies immediately.
Network Data Management: Monitoring and securing massive data flows from numerous IoT
devices. Network analytics tools like NetFlow/IPFIX are used for traffic analysis.

Department of MCA 3
IoT (22MCA32)

Handling Unstructured IoT Data: IoT generates diverse unstructured data (images, audio, video)
that requires advanced techniques like machine learning and computer vision.
Example: A smart city deployment faces challenges in handling the high volume of data from
various IoT sensors, requiring real-time analysis at the edge and scalable data storage solutions.
Additionally, analyzing CCTV footage for security and traffic management involves handling
unstructured video data using computer vision techniques.

Conclusion
In summary, the massive scale of data generated by IoT devices across industries presents both
challenges and opportunities. While traditional analytics approaches face limitations with IoT's
data volumes and varieties, emerging technologies like NoSQL databases, stream processing,
edge computing, and machine learning are enabling organizations to effectively harness IoT data
insights. By overcoming the hurdles of data management and analysis, companies can leverage
IoT analytics to drive operational intelligence, process optimization, and data-driven
decision-making for competitive advantage.

Department of MCA 4

You might also like